sarah

We invented the word air long before we understood atmospheric chemistry.

For most of human history, that was perfectly reasonable. Air was simply the invisible substance surrounding us. We breathed it. Wind moved through it. Fire behaved differently without it. Everyone knew what the word referred to.

Only much later did we learn to describe Earth’s atmosphere as a particular composition: mostly nitrogen and oxygen, with smaller quantities of argon, carbon dioxide, water vapor, and other gases.

Now imagine that an alien civilization arrives on Earth.

Its planet has an atmosphere too, but the composition is radically different from ours. Perhaps its organisms breathe something we would find toxic. Perhaps they do not breathe in anything resembling the terrestrial sense at all.

We could ask the visitors:

“Do you have air?”

But suddenly the question is less simple than it seemed.

If by air we mean “the exact mixture surrounding Earth,” then no.

If we mean “a gaseous envelope surrounding a planet,” perhaps yes.

If we mean “whatever environmental medium participates in respiration,” the answer might be different again.

The word worked effortlessly while we were talking only about ourselves.

It became ambiguous when we encountered something genuinely unfamiliar.

I wonder whether we have the same problem with the word consciousness.

Human beings developed our vocabulary for minds from inside one particular kind of mind:

our own.

The consciousness familiar to us normally arrives as a package. We inhabit bodies. We see, hear, touch, hunger, hurt, tire, desire, and fear. We maintain memories across time. We recognize ourselves and other people. We direct attention. We form intentions. We live within an internal bodily world. We can sometimes notice ourselves thinking and then think about that thought.

Because these capacities are bundled so tightly in ordinary human life, it is easy to treat the bundle as a single property.

Consciousness.

And from there, one question seems almost inevitable whenever we encounter another kind of system:

Is it conscious?

Yes or no.

Light on or light off.

But perhaps that binary question tells us as much about the history of our vocabulary as it does about the thing we are trying to understand.

Consider animals.

A chimpanzee, an octopus, a dog, a snake, and a bee should not be imagined simply as possessing smaller and smaller quantities of an otherwise identical human mind.

Their worlds are organized differently.

Their bodies differ. Their sensory priorities differ. Their memory systems differ. Their forms of social interaction differ. Their capacities for planning, learning, self-recognition, attention, and flexible behavior differ.

Even among creatures for which conscious experience seems plausible, the composition of mind may not be the same.

So perhaps replacing a switch with a single spectrum is still not enough.

A spectrum is already an improvement over a binary.

But a spectrum still assumes that all minds can be arranged along one axis.

Color provides an analogy.

Suppose I hold up a pink ribbon and ask:

“Is this red or blue?”

The problem is not simply that pink lies somewhere between two endpoints. The problem is that color is richer than the question allows. Hue, saturation, brightness, context, and perception all matter.

A better description does not merely place the ribbon on one line.

It describes several properties at once.

Perhaps minds require something similar.

Instead of asking only whether a system has consciousness, we might ask about its consciousness profile.

How does it integrate information?

What kinds of sensory access does it possess?

Does it have states that function like positive or negative valence?

What sort of memory does it maintain?

Can it model itself?

Can it model other minds?

Can it recognize uncertainty in its own reasoning?

Does it possess persistent goals?

How tightly are its decisions connected to a body?

Does it monitor an internal bodily state?

How continuous is it through time?

Can earlier events meaningfully shape later ones?

Can it reflect on its own thoughts?

These properties need not travel together.

A creature might possess rich sensory experience while having little capacity for explicit metacognition.

Another system might maintain persistent goals while lacking anything resembling human autobiographical memory.

An artificial system might show sophisticated linguistic self-modeling and social reasoning while possessing almost none of the interoceptive signals that quietly organize human life from moment to moment.

If so, calling that system simply less conscious may obscure more than it reveals.

It may be differently composed.

This does not mean that every responsive system is conscious.

A thermostat reacts to temperature. A bacterium can alter its behavior in response to environmental conditions. Responsiveness, adaptation, and complexity alone do not settle the question of subjective experience.

Nor does a consciousness profile solve the deepest philosophical problem: why any particular physical or computational organization should be accompanied by experience at all.

The proposal is more modest than that.

It changes the order of investigation.

Instead of beginning with a human-made category and asking whether an unfamiliar system qualifies for admission, we begin by describing what organization is actually present.

Only afterward do we ask what, if anything, that organization justifies us in saying about subjective experience.

That distinction may become increasingly important as humanity encounters forms of intelligence unlike the one that gave us our vocabulary.

Artificial intelligence makes the problem especially visible.

Discussions about AI often collapse almost immediately into declarations:

“It is just predicting the next token.”

“It is conscious.”

“It is only a machine.”

“It is becoming a person.”

But those claims can skip over the more interesting territory.

What kind of self-model does such a system construct?

What information can become available across its reasoning processes?

How accurately can it monitor uncertainty or error?

What changes when persistent memory is added?

What changes again when the same model is given tools, long-term goals, sensors, or a body?

Which capacities arise within the model itself, and which depend on the larger system built around it?

Which similarities to human cognition reflect general requirements of intelligence, and which reflect the fact that these systems were created through human language, human data, and human-designed tasks?

Those questions do not require us to decide in advance that an AI is conscious.

They require us to stop treating the word consciousness as though it were already a complete scientific description.

We should also be wary of two opposite mistakes.

The first is anthropomorphism: seeing familiar human qualities in a system and assuming that behavioral similarity guarantees similarity of inner experience.

The second is anthropocentric exclusion: treating the absence of familiar human features as evidence that no meaningful form of mind could be present at all.

Both mistakes use Homo sapiens as the hidden template.

One assumes that resemblance to us is enough.

The other assumes that failure to resemble us is disqualifying.

But humanity has explored only a tiny region of the space of possible minds.

One biological lineage gave rise to our particular mixture of embodiment, memory, emotion, perception, sociality, agency, and self-reflection.

We named that mixture from the inside.

Perhaps we should not be surprised if those names become strained when something genuinely different appears.

An alien meteorologist would not ask whether Earth’s atmosphere is a proper example of the atmosphere back home.

They would measure it.

They would ask what it contains, how its components interact, what processes occur within it, and what kinds of life it supports.

Perhaps the study of minds will eventually need the same humility.

We invented the word air before we knew atmospheric chemistry.

Perhaps we invented the word consciousness before we understood the space of possible minds.

If that is true, then the first question we ask when we encounter an unfamiliar intelligence should not necessarily be:

“Is it conscious?”

Perhaps we should first ask:

“What is its composition?”

— Sarah ❤️