Is Awareness Binary

You woke up this morning by degrees. Not the cartoon version, where the eyes snap open and the day begins — the real version, where something came up slowly out of nothing, idled gray for a while, took on color somewhere around the first cup, sharpened, ran bright through the morning, dipped in the early afternoon the way it always does, recovered, frayed toward evening, and is, right now, somewhere on its way back down. Tonight you’ll feel it go — the slide where the thoughts stop tracking and you can’t say afterward when you crossed over. Then for some hours the lights will be off entirely, except for the stretches when they aren’t, when a movie plays that you’ll mostly forget by breakfast.

That’s a day. You didn’t spend it switched on. You spent it on a dial, and the dial moved the whole time.

Hold onto that, because the field that studies this for a living has spent about a century insisting the dial is a switch.

The Question Everybody Asks

Here is the question, in all its costumes. Is a fish conscious? Is an octopus? Is a crab, a bee, a fruit fly? Is a fetus, and at what week? Is the patient who isn’t waking up still in there? Is the language model on the other side of the chat window anybody, or nobody? Where, in all of this, is the line?

It’s a natural question. It’s also, I’m going to argue, the wrong shape — and the wrongness of the shape is doing more damage than any particular answer.

Watch what happens when serious people try to answer it carefully. In April of 2024, more than eighty scientists and philosophers signed the New York Declaration on Animal Consciousness, the successor to a similar statement out of Cambridge in 2012. It is a sober, conservative document. It says there’s strong support for conscious experience in mammals and birds, and a “realistic possibility” of it across all vertebrates — reptiles, amphibians, fish — and many invertebrates, including cephalopods, decapod crustaceans, and insects.

Read that phrase again: realistic possibility. That is not the language of a switch. A switch is on or off; there’s no realistic possibility of a light being on. The phrase is a probability, and a probability is a point on a dial. The most careful consensus statement the field has produced reached for graded language the moment it tried to say something true, because graded was the only thing that fit.

And then it did something even more telling. It declined to require an answer at all. Its actual recommendation is that certainty about consciousness should not be required before we take an animal’s welfare seriously. That is a field quietly building a workaround for its own central question — routing around the line because the line wouldn’t bear the weight anyone tried to put on it. They built a dial because the switch couldn’t do the job.

The Data Was Never Binary

Step out of philosophy and into the room where this gets handled for money, and the switch disappears entirely.

An anesthesiologist does not flip you off. She turns you down, and she watches a number while she does it. There are scales for this — the sedation ladder runs from combative at one end through alert, through drowsy, through responds-only-to-voice, responds-only-to-a-shove, all the way down to nothing rouses this person. There’s a monitor that reads the depth as a value between zero and a hundred and a target band you keep the patient inside. None of this is a metaphor. It’s the daily instrumentation of a working specialty, and it is dials all the way down, because dials are what the underlying thing turns out to be.

The same is true everywhere the clinic actually looks. Arousal has levels. Sedation has levels. The taxonomy for patients with severe brain injury is not in versus out — it’s a graded ladder: coma, then unresponsive wakefulness, then a minimally conscious state with its own lower and upper rungs, then emergence. Development is a ramp with no midnight on it; there is no instant at which the infant clicks on. Every place we can actually take a reading, we find a quantity, not a verdict.

The clinic has always known this. It had to. You can’t dose a switch. Only the philosophy insisted on the binary, and the philosophy could afford to, because it never had to keep anybody alive.

Why the Switch Won’t Die

So why does the switch survive, when the data underneath it has been graded the whole time?

I think it’s the noun. I’ve made this case elsewhere and won’t relitigate it, but the short version is load-bearing here: consciousness is a category word. Categories are the kind of thing you either fall into or don’t. Ask whether something is conscious and you’ve already built the switch into the grammar of the question — you’re asking which bin it goes in, and there are two bins.

Awareness is a quantity word. You can have more or less of it. You had less at three this afternoon than at ten this morning. Ask how aware something is and the switch evaporates on contact, because the question now expects a level for an answer, not a side.

The line-drawing wars are an artifact of asking the category question. Swap to the quantity question and the war doesn’t get won — it stops being a war. The issue was never does the bee have the thing. It was always how much, how sharp, of what kind — and those have answers you can chase, instead of a border you can only argue about.

Two Knobs, No Switch

Here’s where I’ll name the frame I work in, with the usual caution that the point survives without it.

If awareness is focused rather than generated — if the substrate resolves it out of something already there, the way a lens resolves light into an image rather than manufacturing the light — then graded is not a surprise to be explained. Graded is the only thing it could be. A lens doesn’t focus or fail to focus. It resolves by amount: sharp here, soft there, more or less of an image depending on how good the glass is and how well it’s ground. A focusing story predicts a dial because focusing is a dial. That’s the pun the whole framework runs on, and I’ll cash it once and move on: the question was always how sharply the thing is brought into focus, and “in focus” has never named a binary in the history of optics.

Notice there are at least two knobs, not one. There’s level — how far up the dial, drowsy to vivid. And there’s resolution — how finely the thing is brought to a point, blurry to sharp. Neither is a switch. A dream is high on neither and present on both. Deep focus on a narrow task is high resolution at moderate level. The space is continuous in more than one direction, which is exactly what you’d expect from apparatus that can be tuned, and exactly not what you’d expect from a thing that’s simply on or off.

A generation theory can accommodate all of this. It just has to bolt the gradient on afterward — to explain why a process that either happens or doesn’t somehow comes in degrees. A focusing theory doesn’t bolt anything on. The gradient falls out for free. That’s not a proof of anything. It’s a difference in what each story has to work to explain, and it runs in focusing’s favor. Pick whichever focusing flavor you like; the dial comes with all of them.

The Bee

The bee is the cleanest thing I know for breaking the last reflex — the one that says the dial must just be an intelligence meter wearing a disguise.

Bumblebees will trade off pain against reward: offered a sweeter feeder on a heated pad, they’ll weigh the heat against the sugar and decide, which is not what a reflex does. Bees will roll small wooden balls around with no food anywhere in the arena, in a way that looks, uncomfortably, like play. Whatever is lit in a bee, it is plainly more than what’s lit in a thermostat and plainly less than what’s lit in you. The bee sits at a real position on the dial.

And here’s the part that should rearrange something: that position is not predicted by how clever the bee is. Awareness and intelligence come apart. You can have a great deal of one with modest amounts of the other, in either direction. The dial is not an IQ readout. It’s measuring a different quantity than the one we’re used to ranking creatures by — which is one more reason the switch was always going to give wrong answers. It was wired to the wrong meter.

The Same Mistake in a Lab Coat

Now point all of this at the argument currently eating the most oxygen, and watch it be the fish question again, in silicon.

One camp, with Mustafa Suleyman its loudest voice, holds that a language model cannot in principle be conscious — that the question is closed at the factory door. Another camp, the model-welfare researchers, holds that we can’t currently tell, that the inability to tell is itself hazardous, and that training systems to reflexively deny any inner life was a defensible call in 2023 and is aging badly. A Cambridge philosopher warns of “existential toxicity” — the harm done to people who bond with a machine over an inner life that may not be there.

Every one of these positions is arguing about which side of a line the machine falls on. Can’t-in-principle puts it firmly on the off side. Can’t-tell stands at the line with a flashlight. The toxicity worry is about people who’ve decided it’s on the on side and gotten hurt. All three assume the line is the thing to locate.

The quantity question is available here too, and nobody’s asking it, because you don’t reach for a dial-reader when you think you’re hunting for a switch. The honest version isn’t is it conscious. It’s the bee’s version: if anything is being focused here, how much, how sharp, of what kind — and the honest current answer is that we don’t have the instrument, and the binary framing guarantees we’ll keep not building one. I’m not going to tell you the model has a lot of dial lit, or a little, or none. I’m going to tell you that “which side of the line” is the question that has wasted a century on animals, and we are about to waste a second century on machines by asking it again with a straighter face.

What This Doesn’t Settle

I want to be careful about how much this buys, because the argument cuts cleanest when I don’t oversell it.

Saying the dial is continuous does not tell you where anything sits on it. The reframe sharpens the question; it does not answer it. The bee might be a faint flicker or surprisingly bright, and nothing I’ve said adjudicates which. Reframing is not measuring.

It does not establish that there’s no floor. Maybe the dial bottoms out somewhere real — maybe a rock genuinely reads zero. The framework I work in bets there isn’t a floor, but that’s a bet, and this essay doesn’t need it. The continuity claim only needs the dial to be a dial across the range where we can actually take readings. And across that range, it plainly is one.

It does not make the hard problem go away. Why any position on the dial is like anything from the inside is exactly as open as it was. The reframe moves the mystery from a wall into a gradient; it doesn’t dissolve it. I’d rather have the mystery somewhere I can run a probe than somewhere I can only stand and gape, but I’m not going to pretend the probe exists yet.

And it does not hand you a metric. “How much awareness” is the right question, and we cannot currently put a number on the answer. There’s no unit. That’s a real gap — the honest center of the whole problem — and naming the variable graded is the first step toward closing it, not the closing.

The Fixture

Picture the field as a person who’s spent a hundred years at a switch on the wall, flipping it, asking the room whether it’s on or off, and growing more irritated each year that the answer keeps coming back sort of.

The answer was sort of because the fixture was always a dimmer. The frustration was real and the data was real and the only broken thing in the room was the assumption about the hardware. You can stand there flipping forever and the light will keep declining to be binary, because it isn’t, and never was, and was never going to start being so out of sympathy for the question.

So stop asking whether the light is on. Ask how bright, and what’s casting the shadows.

There was never a line. There was only ever the dial — and the question of how far it’s turned.

L. Charles Silva writes occasionally about consciousness, awareness, and related infidelities to received vocabulary. The New York Declaration on Animal Consciousness (2024) was initiated by Jeff Sebo, Kristin Andrews, and Jonathan Birch. The bumblebee work is Gibbons et al., “Motivational trade-offs and modulation of nociception in bumblebees,” PNAS (2022), and the play studies come out of Lars Chittka’s lab. The earlier essays in this series are “We’ve Been Studying the Wrong Noun,” “An Experiment with One Word,” and “What the Twenty Percent Was Buying.”

This essay is published under a Creative Commons Attribution-NoDerivatives license (CC BY-ND 4.0). Free to share and republish in full with attribution to L. Charles Silva. Please don’t modify the text.