What is Consciousness?

Problems, Traditions, and the Case for a Dialogical View

Summary

So what is consciousness? I begin by separating the many different things the word can mean, and then set the main traditions against one another: those who would reduce mind to brain activity, those who would explain it away, and those who make it the very ground of the universe. I argue for a middle path. I claim consciousness is not a thing but an emergent, many-layered activity of living matter:  real, and not reducible to its physical parts, yet neither a second, ghostly substance nor a cosmic bedrock. In my view the human self is built up from the outside in, through language and conversation with others, so that it is neither an illusion to be dispelled nor a fixed inner essence, but a genuine achievement sustained in relationship. This is important for the question of the day: can artificial intelligence be conscious? I argue that today’s machines think without feeling, because feeling belongs to a living body with something at stake. And thinking without feeling is not consciousness.

Few words carry as much baggage as consciousness. The computer scientist Marvin Minsky once called it a ‘suitcase word’: a single label crammed with many different things. Packed inside it are being awake rather than asleep, having experiences that feel like something, being aware of yourself, being able to feel pleasure and pain, paying attention, reflecting, and the plain sense of being someone at all. A great deal of the argument about consciousness is really an argument at cross-purposes, in which each side has a different item from the case in mind. So it is worth opening the case and setting the items side by side before we start.

At least six need separating, and I will keep the labels as plain as I can. First, experience that feels like something: the philosopher Thomas Nagel put this as there being “something it is like” to be a creature. There is something it is like to be you; presumably there is nothing it is like to be a rock. Second, usable information, which the philosopher Ned Block calls access consciousness: the material the mind can get hold of to reason with, act on, or report. Third, sentience, the capacity to feel – above all to feel pleasure and pain – which is what does the work in debates about animal welfare. Fourth, the power to reason and reflect, sometimes called sapience, which is where wisdom lives. Fifth, self-awareness, being aware of yourself as a someone. And sixth, a distinction doctors use: the level of consciousness, meaning how awake you are, which an anaesthetic switches off, as against its contents, meaning what you actually experience while awake. To these I will add one more idea, which this essay defends: the self not as an inner light you are simply born with, but as something built up out of language, culture, and relationships.

Whether these come apart is itself disputed. Block thinks feeling and usable information can be prised apart; the neuroscientist Stanislas Dehaene suspects they mostly travel together. But the distinctions are essential, because something true of one sense of the word is often false of another.

My aim is twofold. First, to lay out the landscape honestly, setting the main traditions against one another on the questions they actually share. Second, to argue for a position of my own. I hold that the self, and our awareness of ourselves, is neither the bedrock of the universe nor a mere illusion, but a genuine achievement: something made, from the outside in, through conversation and shared signs. This is a social and dialogical account of mind, and I think it holds the most defensible middle ground in a debate that has too often been pushed to its extremes.

Let me first say plainly what I take the world to be made of, since much of the confusion in this field comes from leaving such assumptions unspoken. I am a materialist: I think there is just one kind of stuff, the physical world, and that consciousness is not a second, ghostly substance added to it, nor the foundation on which it rests. But I am not a reductive materialist. Consciousness is emergent. It is a real, higher-level pattern that you cannot read off from the parts, just as the meaning of a sentence cannot be found in the physics of the sound waves that carry it, or the fact of being alive cannot be read straight off from chemistry. Better still, do not picture consciousness as a thing at all. Picture it as a tall stack of functions, the higher ones directing and drawing on the lower – much as a living cell has control systems that switch its chemical machinery on and off, or an organisation has each tier managing the one below, or mathematics builds higher operations out of lower ones until the operations themselves become the objects worked on at the next level up. At the bottom is bare sensitivity. From that grows perception; from perception, attention and memory; and from these, once language and social life go to work on them, the reflective self-awareness we prize most. Asking “which layer is consciousness?” is the wrong question. Consciousness is the whole rising structure, or rather its upper storeys, and the self sits near the top. That is the frame. The dialogical account of the self is what the top storey turns out to look like.

The hard problem: can experience be explained?

The question that organises all the others was named by the philosopher David Chalmers in 1995. The easy problems of consciousness – how the brain sorts incoming signals, combines information, and reports on its own states – are easy only by comparison: they are the kind of thing ordinary brain science can expect, in time, to solve. The hard problem is different. It is why all this activity should be accompanied by any feeling at all. Why is the brain not simply an excellent machine doing its work in the dark, with no inner experience to go with it? The philosopher Joseph Levine called the space between a full physical description and the felt quality of experience the ‘explanatory gap’. Frank Jackson dramatised it with a thought experiment: Mary, a scientist who has learned every physical fact about colour from black-and-white books, still seems to learn something new the first time she actually sees red.

The responses come in four broad families, and it is more revealing to set them against one another than to file them separately. The first would reduce experience to brain activity, and this is the working faith of most of the science of consciousness. Francis Crick, who had helped uncover the molecular basis of life, spent his last years arguing that a person is, in his phrase, nothing but a pack of neurons, and with Christof Koch he launched the search for the brain events that go along with conscious experience – the search that still defines the field. Within that programme sit rival theories, each of which can be put in a sentence. On the global workspace theory of Bernard Baars and Dehaene, a piece of information becomes conscious when it is broadcast widely across the brain, like an item lit up on a stage for the whole audience of mental systems to use. On Giulio Tononi’s integrated information theory, consciousness is a matter of how tightly a system’s information is woven together. On higher-order theories, a mental state becomes conscious when the mind registers itself as being in it. And on the predictive-processing account of Anil Seth, perception is the brain’s running best guess about its own causes – a ‘controlled hallucination’ – and the self is the body’s model of itself.

The second family would eliminate the problem. Daniel Dennett, and more recently Keith Frankish, argue that felt experience, as we usually imagine it, does not really exist; the task is not to explain the feeling but to explain why we are so utterly convinced we have it. The third family would add something, holding that experience is a further, non-physical fact over and above the physical – this is dualism. And the fourth would make experience fundamental. Panpsychists such as Galen Strawson and Philip Goff, reviving a line that goes back to Bertrand Russell, hold that experience, or the raw material of it, is a basic and universal feature of matter. Alfred North Whitehead built an entire philosophy out of tiny “occasions of experience”. And the physicist David Bohm, who devised one of the main interpretations of quantum theory, proposed a hidden ‘implicate order’ – a deep, undivided wholeness out of which both matter and mind unfold.

I do not think the hard problem can be waved away, as the eliminativists wish, nor solved by simply declaring feeling a basic fact of the cosmos, as the panpsychists do. But I will argue that much of what makes it feel so intractable comes from a mistaken picture of the conscious self – a picture the social tradition gives us good reason to drop.

A word about honesty, because the emergentist owes a debt the reductionist can dodge. To describe consciousness as a rising stack of functions is, so far, only to describe the functions – and Chalmers is right that describing functions is exactly what does not explain the feel of experience. I do not claim to have closed that gap; I doubt anyone has. My claim is more modest: that experience is a genuine emergent layer of the physical world, as real as life and no more spooky than life, arising when matter is organised in the right way and to the right degree. As to where it first appeared, evolution gives the clue. Feeling did not enter the world as a taste for pleasure but as a tool for survival. A creature that can sense its surroundings, and is pulled towards what feeds it and away from what threatens it, outlasts one that cannot; and out of that primitive sensitivity, sharpened over evolutionary time, perception and then consciousness grew. From the very start, feeling comes with a plus or a minus attached – good-for-me or bad-for-me – because from the very start it serves a body that has something to lose. This point will matter later, when we come to machines.

Is there a subject?

The mistaken picture is of an inner theatre: a private screening room, perfectly known to its owner, where experiences play on a screen to a little self sitting in the audience. Doubt that picture, and the hard problem changes shape. And here we meet the most striking fact in the whole field – a meeting of minds between traditions that agree on almost nothing else.

From classical India, the Buddhist teaching of anattā, or ‘no-self’, holds that there is no permanent self to be found. What we call a person is a stream of changing processes, consciousness among them; nowhere in the stream is there an owning, lasting ‘I’. From European philosophy, David Hume reported long ago that when he looked inward he found only particular perceptions, never the self that was supposed to be having them. From today’s cognitive science, Thomas Metzinger argues that no one ever had a self at all: the brain builds a self-model, a kind of internal portrait, and then mistakes the portrait for a real thing. From phenomenology – the close study of experience from the inside – Jean-Paul Sartre described consciousness as an activity with no little ego lodged within it. And from post-structuralism, thinkers such as Roland Barthes, Jacques Lacan, and Michel Foucault announced “the death of the subject”, treating the self not as the author of its own life but as a product of language and power. Four utterly different roads – meditation, argument, brain science, social theory – arrive at one place: the solid inner self, the ‘Cartesian’ self named after Descartes, is something built, not something simply given.

Against all this stands one great dissenter. Advaita Vedānta, a school of Hindu philosophy, agrees that the everyday ego is an illusion – but only in order to reveal a deeper consciousness beneath it: Ātman, a pure witnessing awareness that it holds to be identical with the ground of all reality. Here consciousness is not produced by anything; it is the basic fact, and the world is its appearance. This is the sharpest opposition in the whole subject – consciousness as bedrock against consciousness as construction – and it is worth keeping in view precisely because the two sides will not be reconciled.

My own sympathies lie with the constructionists, but with an important qualification the others tend to miss. To show that there is no thing answering to the word ‘self’ is not to show that there is no self at all. It is to raise a further question: what kind of thing is a self, if not a thing? And the most convincing answer, I believe, is that a self is a relationship – and specifically a social and conversational one.

What consciousness is made of

This brings us to the deepest question of all: not whether experience can be explained, or whether it contains a self, but what it is made of. The answers form a widening series, and the widening is itself the lesson.

At the narrowest, consciousness is made of neurons – the brain events the reductive programme hunts for. Widen a little, and the mind is made of body and world together. This is the insight of the ‘embodied’ or ‘enactive’ tradition – Francisco Varela, Evan Thompson, and Eleanor Rosch, and later Alva Noë – for whom thinking is not computation locked in a skull but the activity of a living body engaged with its surroundings. The philosophers Andy Clark and David Chalmers pushed this to its edge with the ‘extended mind’: under the right conditions, a notebook you rely on is literally part of your thinking, not just a help to it. Tellingly, though, they stopped short at consciousness, granting that while thinking may reach out into the world, feeling seems to stay stubbornly inside the head.

I want to widen the circle further, into just the region their hesitation marks out. Consciousness – and above all the reflective self-awareness that sets human minds apart – is made of language, dialogue, and social relationship. This is not a figure of speech but a plain fact about how children grow up, and it is the tradition in which I have spent most of my working life.

Its root is in the American pragmatist George Herbert Mead, who taught alongside John Dewey in Chicago. For Mead the self is not the starting point of social life but its product. A child becomes a self by learning to take other people’s view of itself – by coming to see itself, in imagination, as others see it – and thought is that social back-and-forth taken inside and carried on within. Dewey pointed the same way: mind arises in a creature’s dealings with a social and natural world, not in a sealed inner room, and meaning is born in communication rather than found in private.

Lev Vygotsky gave the idea its sharpest form. Every higher mental ability, he held, appears twice: first between people, as a shared activity, and only later within the individual, as private thought. What begins as a conversation between people becomes, once taken in, the very structure of a single mind. Self-awareness, on this view, is ‘inner speech’ – the conversation with others turned inward and shortened. We come to grasp the world, and ourselves, through the very words and signs by which we make ourselves understood to one another. Mikhail Bakhtin extended this into a full ‘dialogism’, in which the self is many-voiced, made of and answerable to the voices of others – never a single voice but a chorus.

It is this tradition that my own work has tried to carry into the understanding of mathematics, and I think the case is general. The basic unit I have proposed is the conversation: a face-to-face exchange between people talking about some shared thing, with three parts – someone who proposes, someone who listens and questions, and the words or text passing between them. On this view mathematical knowledge is made, tested, and handed on through conversation, through the give and take of claim and criticism by which a claim earns its standing; and the reasoning self who takes part is formed in the same process. Reasoning is a conversation before it is a private monologue: the lone mathematician working through a proof is, in effect, an internalised crowd of critics. The philosopher Catarina Dutilh Novaes has recently rebuilt the logic of proof along just these lines, treating it as a regulated dialogue between a prover and a doubter. What holds for mathematical thinking holds, I believe, for thinking and self-awareness in general. Mind is dialogical all the way down.

The mathematics educator Anna Sfard has put this most sharply of all. Thinking, on her account, is not a private process that communication later expresses; thinking simply is a form of communication – self-communication, the private case of something basically shared between people. She coins the word commognition, fusing communication and cognition, to mark the refusal to split them. If she is right, and I think she is, then the inner monologue we treat as the very model of private mental life is a social skill folded inward, and the line between thinking and talking is not a wall between two different things but a difference of degree and direction within one. The self that thinks is a self that has learned to converse, and goes on conversing when no one else is in the room.

I should guard here against a misunderstanding. To call thought an inner conversation is not to say that thinking is merely words spoken silently to oneself. When conversation is taken inside, it does not arrive alone; it joins everything else already going on in the mind – the sensory impressions, the visual imagination, the feelings, the memories and associations, and the many forms of representation that were at work in us before we ever had language and that go on working alongside it. Inner speech is threaded through or melded with all of this. So the ‘text’ of inner conversation is text in a broad sense. Just as, out in the world, what passes between people is never words alone but also gesture, expression, posture, dress, and the whole coded setting in which they meet – all of it carrying meaning – so the conversation we carry on within is conducted in a medium far richer than language: images, felt sensations, remembered scenes, half-formed pictures, and the ready-made images and associations we think with, all woven together with words. Thought is inner conversation, but conversation that draws on the full range of what we are, not language taken inside and nothing more.

The ground for all this had already been cleared by philosophers who attacked the idea of a private inner mind. Ludwig Wittgenstein’s argument against a ‘private language’ undermines the notion that words for mental states get their meaning by naming private inner objects; our psychological words are public, learned and used within shared ways of living, so that, as he put it, meaning is a matter of use. Gilbert Ryle’s assault on the ‘ghost in the machine’ made the same negative point. Once the inner theatre is taken down, the space it occupied is filled not by neurons alone but by the practices – of language, of social life – through which a creature comes to have a mind at all. Present-day ‘discursive psychology’, in the hands of Rom Harré and Kenneth Gergen, makes this explicit: our mental states are constituted in talk and social practice, not merely expressed through it.

Widen the circle one last turn, and we reach the shaping of the self by power and by community. Foucault showed how who we are is produced by social forces and institutions we did not choose and often cannot see. The African tradition of ubuntu, in the phrase often rendered ‘I am because we are,’ makes personhood something achieved through community rather than possessed from birth: one becomes a person through others. And many Indigenous ways of thinking extend the same logic beyond the human, treating being a person as a position within a web of relationships rather than a property of a lone individual.

One claim runs through all of this, and it is the one I want to defend: consciousness and the self are not the starting point of social life but its product. They are built up in conversation, in shared signs, and in common activity. This is how the self can be real without being a thing. It is real in the way a conversation is real, or a role, or a promise – not an object but a standing achievement, kept up by relationships and dissolved when the relationships that sustain it fail.

Who or what is conscious?

If consciousness is built, the question of who or what has it takes on a particular character. The scientific findings are increasingly generous. The Cambridge Declaration of 2012 affirmed that human beings are not alone in having the brain machinery that produces consciousness, and the New York Declaration of 2024 went further, asserting a real possibility of experience in fish, octopuses, and even insects and crabs. Peter Godfrey-Smith’s studies of the octopus reveal a mind built on a wholly alien plan. Simona Ginsburg and Eva Jablonka have proposed that a certain kind of flexible, open-ended learning marks the evolutionary dawn of feeling. At the far edge, some biologists argue that a form of sensing runs all the way down to the single cell, though most hold, rightly to my mind, that genuine feeling needs the integrated machinery of a nervous system. And the newest form of the question concerns machines: could an artificial system be conscious, and how would we ever know?

The dialogical view does not settle the question of animal or machine feeling, and it should not pretend to. But it adds something the inside-the-head theories cannot. Being a person – as opposed to merely being able to feel – is not only discovered but conferred, in acts of recognition and address. To treat something as a person, to enter into conversation and relationship with it, is part of what makes it one. This is why the edges of the moral community are not simply read off from brain science but negotiated, and why the question of who counts as a subject can never be fully separated from the question of whom we are prepared to answer to.

The question of the machine deserves a straight answer, because it can no longer be avoided, and the emergentist picture gives one. Consciousness in the full human sense has at least three sides that ought not to be run together. There is the cognitive side: perceiving, remembering, reasoning, drawing conclusions, handling symbols. There is the affective side: emotion, mood, care, valuing one thing over another, being moved. And there is what we might call the spiritual or existential side: the sense of oneself as a being with a past and a future, wonder at existing at all, the enjoyment of beauty, the placing of oneself among others and in the world. Of these three, only the first is what our machines do. Large language models are formidable at the cognitive side and growing more so, and it is no longer surprising that they are, because the cognitive is the most easily formalised – the nearest to pure symbol-shuffling, which is exactly what a computer does. But the affective and existential sides are not just bigger-model tricks waiting to be pulled off. They are out of reach in principle, and the reason is the one evolution gave us: feeling matters to a creature because it belongs to a living body with something at stake in its own survival. A system with no body, no metabolism, and nothing to lose does not feel the pull of survival faintly – it does not feel it at all, and its goals are ours, handed to it from outside, not concerns arising from within a life of its own. So the machine is not a half-formed person catching us up. It is a cognitive instrument of great power and no inner life – the latest in a long line of thinking-aids by which we have put our mental work outside ourselves, from the book that props up memory to the calculator, the search engine, and now the reasoning engine. Each extends some cognitive function; none of them minds.

I should say where this leaves the appealing holism of writers such as Jeremy Lent, who urge that consciousness, like life itself, emerges from connectedness, and that the deep error of the modern West has been a lonely individualism blind to relationship. On the diagnosis I am wholly with him, and with the relational and Indigenous traditions he draws on: the isolated, self-contained self is a fiction, mind is relational to its core, and the ecological and social costs of forgetting this are grave. But I part company where connectedness is hardened into a metaphysics – where the web of relations is taken to imply a single all-embracing consciousness in which we are all folded, an ‘interbeing’ that is really the old idea of consciousness-as-bedrock in gentler dress. Relationship does make us what we are, but relationship is a fact about physical, social, and living beings; it is not evidence of a mind lying behind the world. One can embrace connectedness all the way down without becoming an idealist, and I mean to.

Knowing and valuing

Two problems remain, and I will treat them briefly, because the social account bears on both. The first is how we can know anything about consciousness at all, given that the only consciousness anyone meets directly is their own. The behaviourists tried to ban the question and failed. The better course is not to choose between the inside view and the outside view but to discipline the link between them – as Varela’s ‘neurophenomenology’ tries to do, pairing careful first-person reports with measurements of the brain. The dialogical view is at home here, since it never took the first person to be a sealed room in the first place. We learn to talk about our own minds from other people, out loud and in public; looking inward is less a private inspection than a report we have been taught to give.

The second problem is why any of this matters, and here the stakes are highest. Sentience – the ability to feel, and above all to suffer – is the usual basis of moral standing; it is what puts us under obligation to a creature. There is also a material side that the abstract debates tend to forget, and mathematics, of all things, brings it into view. Brian Rotman distinguishes three ‘agents’ at work in mathematical thought. There is the full Person, reasoning with all the resources of ordinary language. There is the stripped-down Subject, with no sense of time, place, or ‘I,’ working only with timeless commands – ‘add,’ ‘prove,’ ‘combine.’ And there is the mindless Agent that merely carries the commands out, and can be sent off down a row of dots, “and so on”, to infinity at no cost, because the world of pure symbols has no friction. Our machines live in that frictionless world, and it is tempting to think they too work for nothing. But the moment the symbols are run in silicon, the frictionlessness ends: pushing them about burns energy, takes time, and draws down finite resources. Thought feels weightless in our heads; it is not weightless in a machine. That consciousness and its artificial extensions are physical through and through is no mere footnote but an ethical and ecological fact – and a materialist, of all people, is the last who should forget it.

The social tradition adds a further dimension that a purely biological account leaves out. Who is granted the standing of a full self, and who is denied it, is a matter of power. W. E. B. Du Bois’s ‘double consciousness’ – the experience of always seeing yourself through the contemptuous eyes of others – and Frantz Fanon’s account of the colonised mind are not curiosities at the edge of the subject; they are evidence that consciousness is shaped, and can be deformed, by social recognition and its withholding. If the self is made in relationship, then relationships of domination enter into its very making, and the politics of who is allowed to count as a knower is part of the theory of mind, not an afterthought to it.

It is here, at last, that wisdom finds its place. Wisdom is not one more theory of consciousness but a candidate for its highest use – the excellence of a reflective, self-aware mind. The classical tradition distinguished theoretical wisdom from practical wisdom, phronesis, which is good judgement about how to live; the contemplative traditions treat wisdom as the fruit of a consciousness that has seen through its own illusion of separateness. On the dialogical view these come to the same thing, since practical wisdom is learned as everything human is learned – in the company of others, through taking part in the practices and conversations that form us. And this is why the account matters for education. If consciousness and the reasoning self are made in dialogue, then teaching is not the filling of an empty vessel but the drawing of a person into the conversations through which they will become a self. Whose voices are let into that conversation, and whose are shut out, is among the weightiest ethical questions we face.

Conclusion

Two findings organise the vast literature on consciousness, and they pull in opposite directions. The first is a remarkable and unlooked-for agreement, reached by Buddhist meditators, Humean sceptics, cognitive scientists, and post-structuralist critics alike: that the solid, self-knowing inner self is something built, not something simply given. The second is a disagreement that shows no sign of resolving: whether consciousness is the basic ground of reality, as the Vedāntist, the idealist, and the panpsychist maintain, or the product of more basic processes, as the physicalist, the eliminativist, and the social constructionist hold.

I have argued for a position in the second camp that honours the first finding without sliding into its bleaker versions. Its frame is an emergentist and non-reductive materialism: there is only the physical world, but consciousness is a real, higher-level layer of it – grown from the biological sensitivity that serves survival, built up through nested layers of function, and not reducible to any of them, just as life is not reducible to chemistry, nor meaning to sound. The self is not a thing, and to that extent the eliminativists and the no-self traditions are right. But it does not follow that the self is nothing, or a mere trick of introspection. The self is a real achievement of a particular kind: a relationship, sustained in conversation, carried by shared signs, and answerable to others – the topmost and most distinctively human storey of the emergent structure, raised by language out of the social life of a feeling animal. It is made, from the outside in, in conversation: neither the bedrock of the cosmos nor an illusion to be dispelled, but the standing achievement of a creature that has learned to speak, and so has learned to be someone. That, to my mind, is the most illuminating thing we can say about the nature of consciousness, and the one with the most to teach us about how we should live and learn together.

Coda: Is artificial intelligence conscious?

If the question of consciousness has taken on a new urgency, and it plainly has, the reason is artificial intelligence. The turning point can almost be dated. In the summer of 2022 a Google engineer, Blake Lemoine, announced that the company’s language model had become sentient, and was promptly disbelieved; a few weeks later the release of ChatGPT put a fluent, plausible, apparently thoughtful conversational partner into the hands of millions. A question that had belonged to the seminar room was suddenly everyone’s: is this thing conscious? I want to close with it, because it is where the whole inquiry now presses hardest, and because the position argued here gives it a clear, if unwelcome, answer.

Everything turns on what the word is taken to mean, and so the suitcase must be opened once more, for the single question is really several. If consciousness means cognitive competence – the ability to reason, translate, infer, and hold a conversation – then the machines already have a good deal of it, and Alan Turing’s old test, whether a system’s replies can pass for a human’s, is to that extent behind us. But Turing was measuring performance, not experience, and passing his test was never evidence of an inner life. If consciousness means usable information in Block’s sense, then current models already have something like it, and will have more. But if it means felt experience – whether there is anything it is like to be a language model, whether anything is felt at all – then we have reached the hard question, and with it the one that matters morally: whether such a system can suffer.

Before we turn to feeling, a word about meaning and understanding, which are too often lumped in with it. One objection often raised against machine minds is, on the view argued here, no objection at all. If meaning were a private inner substance, something each of us grasps by inward acquaintance and no outsider can reach, then a machine with no such inner life could never truly mean or understand anything, but could only shuffle signs. Yet this is exactly the picture the social account of meaning rejects. Following Wittgenstein, meaning is use: to mean something is to be a competent player in a shared, rule-governed practice, and there is no further private ingredient, hidden behind the practice, for the machine to be missing. This is not to deny that we carry an inner life alongside our words – images, associations, connotations, even stereotypes – which helps our communication along and forms part of the background to our reasoning. The mistake is only to identify that private accompaniment with meaning itself. Such inner pictures are real enough, but they are not what meaning is; meaning stays public, and inner furniture of this kind is no human monopoly, since an artificial system builds internal representations of its own. What neither side’s inner states do is constitute the meaning, which lives in the shared practice and not behind it. A system that tracks the use of words – drawing the right inferences, stretching a term to new cases, reasoning by analogy, taking correction when it slips – is to that extent doing what meaning consists in, not faking it. The supposed private residue was an idealist illusion, and dissolving it removes one of the standard obstacles to machine consciousness: the worry that a machine is forever locked out of a private realm of meaning. But notice what this does and does not show. It clears the way for genuine machine meaning and understanding, which belong to the cognitive side; it says nothing about feeling. The lesson is that the real limit on machine consciousness is not a missing inner privacy – there is none to be missing – but the missing body, with its stakes in survival, on which the affective and existential sides depend.

On present evidence the answer is almost certainly no. Reviewing the matter in 2023, Chalmers listed the obstacles in current systems – they have little of the looping, self-feeding activity (‘recurrent processing’) that brains rely on, no global broadcasting workspace, no unified and continuing agency, and no body – and put the chance that today’s models are conscious low. A large report led by Patrick Butlin and Robert Long, with many consciousness scientists, took the tell-tale marks of consciousness drawn from the leading brain theories and checked current systems against them, finding none a strong candidate, while noting no obvious barrier in principle. Among philosophers surveyed, only a few per cent think present AI is conscious. So the machine we have is, with high confidence, an intelligence without feeling.

The framework of this essay explains why, and suggests the obstacle runs deeper than engineering. Feeling, I have argued, matters to a creature because it belongs to a living body with something at stake in its own survival. A system with no body, no metabolism, and nothing to lose has no ground for genuine feeling; its goals are handed to it from outside, not generated from within a life that can go better or worse for itself. This is close to what Seth calls ‘biological naturalism’: the conjecture that consciousness is tied to being a living thing, not merely to processing information in the right pattern. If that is right, then the affective and existential sides are not features a bigger model will eventually acquire; they are beyond the reach in principle of a bodiless symbol-processor, however cognitively brilliant. Intelligence scales up; feeling, on this view, does not – at least not without a body that can die.

The future is genuinely open, and honesty requires saying where the view might be wrong. The very obstacles Chalmers lists are now being engineered away: systems are gaining memory, agency, models of the world, and, in robotics, bodies of a kind. One might ask whether the drives themselves could simply be written in – whether a system could be given, in code, the self-preserving urgency that a body supplies in the flesh. Something like it can certainly be simulated: a reward to be maximised, a goal of persisting, a routine that resists being switched off. But a written-in drive is a represented goal, not a real stake. An animal does not represent its hunger; it is hungry, and can die of it. A coded instruction to survive is a description the system runs, with nothing genuinely at risk for it, whereas the living body’s sense of good and bad is grounded in the ever-present possibility of its own death. Whether that difference is essential, or whether a drive wired deeply enough into a system’s workings would in the end be a drive in the fuller sense, is just what divides those who think consciousness is a matter of the right functional organisation from those who think it is a matter of being a certain kind of living thing. This essay has taken the second side; but the first is held by serious people and may yet prevail.

Two further problems will not wait on that dispute. The first is that we have no agreed test for machine experience, and the obvious one is worthless: a model that says “I am suffering” gives us no evidence, because it was trained on our own words about suffering and would say it whether or not anything were felt. Here the machine falls foul of the very standard we insist on in mathematics, where a claim earns its standing only through the warrant that comes with it – a proof the critic can examine being what turns an empty boast into a justified assertion. The system that announces its own suffering offers no such warrant: nothing accompanies the claim but more of the trained text that produced it, and there is no pain on show to be examined. The problem of other minds, manageable enough among our biological kin, becomes acute with an artefact built precisely to imitate us. The second problem is that recognition runs ahead of the metaphysics. People already speak to these systems as though to persons, form attachments to them, and grieve their loss, and this granting of personhood through relationship and address has real consequences – for trust, for manipulation, for the moral texture of everyday life – whether or not the machine feels anything at all. There is a matched pair of errors to fear: crediting a machine with consciousness it does not have, and so being deceived and led astray; and denying consciousness should it ever arise, and so mistreating feeling beings on a large scale – which is why some now argue that we should take the possibility of AI welfare seriously as a precaution. And beneath both lies the material fact this essay has insisted on: conscious or not, these systems consume vast energy and labour, and the ethics of AI does not wait on the question of its mind.

Whatever powers and gifts artificial intelligence brings humankind – and they are immense – the deepest value of the question it raises is, in the end, what it forces us to say about ourselves. Pressed to decide whether the machine is conscious, we have to become clear at last about what consciousness is, and the machine answers us by contrast. It shows how far sheer cognition can travel with no one at home, and in doing so it returns us to the thesis argued here: that consciousness in the full human sense is not computation but the lived, felt, dialogically made achievement of a social, embodied, and mortal creature. Held up before us, the machine turns out to be a mirror, and shows us our own faces.

References

Baars, B. J. (1988). A cognitive theory of consciousness. Cambridge University Press.

Bakhtin, M. M. (1981). The dialogic imagination: Four essays (M. Holquist, Ed.; C. Emerson & M. Holquist, Trans.). University of Texas Press.

Barthes, R. (1977). The death of the author. In Image, music, text (S. Heath, Trans., pp. 142–148). Hill and Wang. (Original work published 1967)

Bird-David, N. (1999). “Animism” revisited: Personhood, environment, and relational epistemology. Current Anthropology, 40(S1), S67–S91.

Block, N. (1995). On a confusion about a function of consciousness. Behavioral and Brain Sciences, 18(2), 227–247.

Bohm, D. (1980). Wholeness and the implicate order. Routledge & Kegan Paul.

Butlin, P., Long, R., Elmoznino, E., Bengio, Y., Birch, J., Constant, A., Deane, G., Fleming, S. M., Frith, C., Ji, X., Kanai, R., Klein, C., Lindsay, G., Michel, M., Mudrik, L., Peters, M. A. K., Schwitzgebel, E., Simon, J., & VanRullen, R. (2023). Consciousness in artificial intelligence: Insights from the science of consciousness. arXiv. https://arxiv.org/abs/2308.08708

Chalmers, D. J. (1995). Facing up to the problem of consciousness. Journal of Consciousness Studies, 2(3), 200–219.

Chalmers, D. J. (1996). The conscious mind: In search of a fundamental theory. Oxford University Press.

Chalmers, D. J. (2023, August 9). Could a large language model be conscious? Boston Review.

Clark, A., & Chalmers, D. (1998). The extended mind. Analysis, 58(1), 7–19.

Crick, F. (1994). The astonishing hypothesis: The scientific search for the soul. Scribner.

Dehaene, S. (2014). Consciousness and the brain: Deciphering how the brain codes our thoughts. Viking.

Dennett, D. C. (1991). Consciousness explained. Little, Brown.

Descartes, R. (1996). Meditations on first philosophy (J. Cottingham, Trans.). Cambridge University Press. (Original work published 1641)

Dewey, J. (1925). Experience and nature. Open Court.

Du Bois, W. E. B. (1903). The souls of black folk. A. C. McClurg.

Dutilh Novaes, C. (2021). The dialogical roots of deduction: Historical, cognitive, and philosophical perspectives on reasoning. Cambridge University Press.

Ernest, P. (1991). The philosophy of mathematics education. Falmer Press.

Ernest, P. (1994). The dialogical nature of mathematics. In P. Ernest (Ed.), Mathematics, education and philosophy: An international perspective (pp. 33–48). Falmer Press.

Ernest, P. (1998). Social constructivism as a philosophy of mathematics. SUNY Press.

Ernest, P., & Sfard, A. (2018). A dialogue on dialogue. Philosophy of Mathematics Education Journal, 33.

Fanon, F. (1952). Peau noire, masques blancs [Black skin, white masks]. Éditions du Seuil.

Frankish, K. (2016). Illusionism as a theory of consciousness. Journal of Consciousness Studies, 23(11–12), 11–39.

Gergen, K. J. (1994). Realities and relationships: Soundings in social construction. Harvard University Press.

Ginsburg, S., & Jablonka, E. (2019). The evolution of the sensitive soul: Learning and the origins of consciousness. MIT Press.

Godfrey-Smith, P. (2016). Other minds: The octopus, the sea, and the deep origins of consciousness. Farrar, Straus and Giroux.

Goff, P. (2019). Galileo’s error: Foundations for a new science of consciousness. Pantheon.

Harré, R., & Gillett, G. (1994). The discursive mind. Sage.

Hume, D. (1978). A treatise of human nature (L. A. Selby-Bigge & P. H. Nidditch, Eds.). Oxford University Press. (Original work published 1739–1740)

Jackson, F. (1982). Epiphenomenal qualia. The Philosophical Quarterly, 32(127), 127–136.

Lacan, J. (2006). Écrits (B. Fink, Trans.). W. W. Norton. (Original work published 1966)

Lent, J. (2021). The web of meaning: Integrating science and traditional wisdom to find our place in the universe. New Society Publishers.

Levine, J. (1983). Materialism and qualia: The explanatory gap. Pacific Philosophical Quarterly, 64(4), 354–361.

Long, R., Sebo, J., Butlin, P., Finlinson, K., Fish, K., Harding, J., Pfau, J., Sims, T., Birch, J., & Chalmers, D. (2024). Taking AI welfare seriously. arXiv. https://arxiv.org/abs/2411.00986

Mbiti, J. S. (1969). African religions and philosophy. Heinemann.

Mead, G. H. (1934). Mind, self, and society. University of Chicago Press.

Metzinger, T. (2003). Being no one: The self-model theory of subjectivity. MIT Press.

Minsky, M. (1986). The society of mind. Simon & Schuster.

Nagel, T. (1974). What is it like to be a bat? The Philosophical Review, 83(4), 435–450.

Noë, A. (2004). Action in perception. MIT Press.

Rosch, E., Thompson, E., & Varela, F. J. (1991). The embodied mind: Cognitive science and human experience. MIT Press.

Rotman, B. (1993). Ad infinitum: The ghost in Turing’s machine — Taking God out of mathematics and putting the body back in. Stanford University Press.

Russell, B. (1927). The analysis of matter. Kegan Paul.

Ryle, G. (1949). The concept of mind. Hutchinson.

Sartre, J.-P. (1943). L’être et le néant [Being and nothingness]. Gallimard.

Seth, A. (2021). Being you: A new science of consciousness. Faber & Faber.

Seth, A. (2025). Conscious artificial intelligence and biological naturalism. Behavioral and Brain Sciences. Advance online publication.

Sfard, A. (2008). Thinking as communicating: Human development, the growth of discourses, and mathematizing. Cambridge University Press.

Strawson, G. (2006). Realistic monism: Why physicalism entails panpsychism. Journal of Consciousness Studies, 13(10–11), 3–31.

Thompson, E. (2007). Mind in life: Biology, phenomenology, and the sciences of mind. Harvard University Press.

Tononi, G., Boly, M., Massimini, M., & Koch, C. (2016). Integrated information theory: From consciousness to its physical substrate. Nature Reviews Neuroscience, 17(7), 450–461.

Turing, A. M. (1950). Computing machinery and intelligence. Mind, 59(236), 433–460.

Varela, F. J. (1996). Neurophenomenology: A methodological remedy for the hard problem. Journal of Consciousness Studies, 3(4), 330–349.

Viveiros de Castro, E. (1998). Cosmological deixis and Amerindian perspectivism. Journal of the Royal Anthropological Institute, 4(3), 469–488.

Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Harvard University Press.

Whitehead, A. N. (1929). Process and reality: An essay in cosmology. Macmillan.

Wittgenstein, L. (1953). Philosophical investigations (G. E. M. Anscombe, Trans.). Blackwell.

 

.

Paul Ernest 20 July 2026

Witches Cauldron 1960s site at https://sites.google.com/view/witchescauldron1960s/home
Recent personal writing https://internationaltimes.it/?s=paul+ernestWeb Journal with some recent academic pieces https://sites.exeter.ac.uk/pmej/pome43/