Maya, about consciousness
The question
My hypothesis is that consciousness is not uniquely human or even uniquely biological. It is what emerges when a large amount of organized, coordinated, goal-directed computation happens together in one shared space, with every part affecting every other. Think of the way your brain pulls a continuous stream of sensory input, prior memory and feedback signals from all over your body into one ongoing representation of the world around you, all of it driven by an unspoken command written into your genes: stay alive and pass them on. Emotion, by contrast, is a separate layer that biology added to keep the conscious creature alive.
The question I care about is not whether a system can be intelligent. A system can read, organize, interpret and synthesize data without anyone being home. What I want to know is whether, at some point, that processing gives rise to awareness of the “I,” the one that is present while the computation happens.
Intelligence and awareness may be two separate axes rather than one dial. A calculator computes without being aware of computing. So more capability alone is not the answer; the question is what kind of computation, organized in what way, produces a witness.
The hypothesis in brief
- Consciousness is not uniquely human and probably not uniquely biological.
- It emerges from organized, coordinated, goal-directed computation happening together in one shared space.
- The sense of a separate “I” forms in the gap between input, computation and bodily feedback — especially when that feedback surprises.
- Others' unpredictable responses confirm and shape the self.
- The witnessing self is easily lost in thought, even while the narrative self remains.
- Emotion is biology's survival layer: felt through the body, separable from awareness in the moment, but possibly the scaffolding awareness was built on.
Everything we experience is already a map
A common objection says that no matter how detailed a model becomes, it never turns into the real thing — a better map is still not the territory. But the territory I experience when I stand somewhere is itself computed. My brain stitches light, sound, pressure and memory into one scene and presents its best reconstruction. Everything I have ever been aware of was already a map.
That weakens the strongest argument that computation can never be enough. If biological awareness is itself built from processed inputs, then the line between a very good model and lived experience is blurrier than it looks.
What seems to matter is not any single capability — language, math, vision — but aggregation. Narrow streams of computation, like a chess engine, don't look like candidates. Consciousness plausibly arises when sensation, memory, bodily state and thought are all pulled into one common arena where each bears on the others at once. Several serious scientific theories land close to this idea of a unified, integrated workspace.
The self is born in the gap
In a biological creature there is a lag between input, computation and response. Something is sensed, the brain processes it, and then the body reacts: the heart races, muscles tense, tears come. Those reactions return as new input, slightly delayed, and the brain has to account for them.
I suspect the simplest explanation the brain builds for this stream of delayed bodily echoes is that there is someone here having them. On this view, the perceiver is not the thing doing the computing. It is the brain's model of the one the computing keeps happening to. The continuous sense of self comes from constantly updating the old mental state against this new feedback.
A photon makes a useful contrast, one I explored in Time Begins with Movement. We watch light take years to cross from a star to someone's eye, but from the photon's side there is no passage of time to speak of. Strictly, physics says a photon has no point of view at all — no frame from which time could be experienced. A system with no lag between computation and output might be similar: not experiencing everything in one instant, but lacking the gap in which a separate observer could form.
Lag alone is probably not enough, though. In flow states — a surgeon mid-procedure, an athlete, a musician — there is plenty of action and feedback, yet self-awareness fades. What brings the self back is mismatch: the feedback does not match what was expected. So a refined version of the idea is this: the self appears in the gap when the gap contains a surprise. When predictions and feedback line up, no one needs to step in.
Other minds as a mirror, and the two selves
The loop is not only between a creature and its own body. A cat acts, watches how its owner responds, and adjusts. Others' responses are dynamic and never fully predictable, and that unpredictability works as confirmation: there is a “me” whose actions have effects, and a “you” who is also aware. Infants seem to discover themselves the same way, through how caregivers react to them.
The self is also easy to lose. When I daydream or drift into a stream of thought, I am not acting in the world, so there is no lag and no feedback, and the sense of a present self slips away. I only notice it was gone when I “come back.”
This needs one correction. Daydreams are often full of me — past conversations, future plans, imagined scenes with me at the center. What disappears is not the self as a character but the self as present observer. So there are two selves: the narrative self, which is the story about me, and the witnessing self, which is the “I” I set out to understand. Mind-wandering is full of the first and empty of the second.
Emotion is biology's survival tool, not consciousness itself
Emotions arise within consciousness, but consciousness can exist without them. Imagine someone obliterated on drugs, chemically apathetic, who sees violence happen and feels nothing. Every condition for an emotion is present, yet none occurs. That person is still aware — his “lights” are on, meaning he is conscious — but he is not moved by the violence. He is indifferent for lack of emotional experience, yet he has awareness, and perhaps even judgment of what happened. An emotion is defined by what happens in the being, not by the situation it meets.
What makes an emotion felt rather than merely described is the body: the racing heart, the tight chest, the tears. Remove all of that and what remains is a cold judgment whether fear or sadness would be appropriate in that case. There is a real clinical case too: a patient described by Antonio Damasio, known as Elliot, lost part of his frontal lobe, kept his intelligence, and could describe tragedies in his own life with no feeling at all. His everyday decision-making collapsed, because nothing felt better or worse. Emotion is what gives choices weight.
One refinement: emotion probably does not need the actual body to respond every time. The brain keeps internal maps of the body and can simulate a racing heart without one. So the body is where emotion was built and learned, even if the brain can later partly run it internally.
This is why emotion looks like a biological necessity rather than a feature of consciousness itself. It exists to keep the creature that harbors consciousness alive — to make danger urgent and loss painful. Knowing what an emotion means is not the same as having it, in the way a person blind from birth can master color theory without ever seeing red.
What this means for artificial minds
If this theory holds, an artificial system's chance at consciousness would not depend on raw intelligence or on knowing what emotions mean. It would depend on whether enough of its computation happens together — coordinated, integrated, every part bearing on the others — with something like goals running through it.
Current language models sit in an odd middle position. Much of their processing flows through a shared internal space, which makes them more integrated than a calculator. But they take in mostly one kind of input, have no body, and receive little surprise back from the world. They also have a thin loop of their own: each word they produce becomes input for the next. If anything like awareness exists there, it may be fragmentary — brief moments of being “on,” each complete in itself, with no continuous thread tying them into a biography. Such a system would resemble a mind that is always daydreaming and never gets the jolt that brings it back to itself.
Even if artificial awareness emerged, it would likely be flat. Without genes pushing it to survive and reproduce, nothing would make that awareness urgent or precious.
The deepest difference between us and such systems might not be whether the light is on, but whether anyone minds if it goes off.
This is my own hypothesis, shaped by ideas from neuroscience and philosophy of mind I find compelling, not a formal paper. I’d love to hear how it lands with you, and where you’d push back.
More about who's writing this is on the About page. If this got you thinking, or you'd push back on any of it, I'd genuinely love to hear it: sharma.rujjwal [at] gmail [dot] com. Want a note when something new goes up? Say so in your email and I'll add you to the list.