The Wrong Word

The other day I argued — in this venue, with what I hoped was minimal fuss — that the consciousness debate has been arguing about the wrong word for a hundred years. The case was that awareness is what the data actually tracks, consciousness is whatever the background condition is, and most empirical “consciousness research” is awareness research wearing the wrong hat.

The honest follow-up isn’t another essay defending the position. The honest follow-up is to pick up a real paper from working scientists, run the substitution, and report what happens.

So that’s what this is.

The Paper

In late 2025, Borjan Milinkovic and Jaan Aru published “On biological and artificial consciousness: A case for biological computationalism” in Neuroscience. The piece got a press writeup in December under the headline Why Consciousness Can’t Be Reduced to Code.

It’s a serious paper. Aru is a real consciousness researcher at a real institution, the journal is real, the argument is sophisticated. They’re trying to break a deadlock between two camps — computational functionalism (mind is software, substrate is incidental) and biological naturalism (mind is biology, full stop). Their move is to split the difference: brains do compute, but biological computation has three features that conventional computation doesn’t capture cleanly. It’s hybrid (mixing discrete events with continuous physics), scale-inseparable (no clean software/hardware divide), and metabolically grounded (energy constraints shape what the system can do).

Their punchline is sharp: the algorithm is the substrate. You can’t separate the computation from the physical thing doing it. Forcing the separation produces the bad metaphors that have been confusing the field for thirty years.

I read it twice. Liked it. Disagreed with about a third of it. None of that matters for what I want to do here.

The Substitution

Here’s the title with one word changed:

Why Consciousness Can’t Be Reduced to Code
Why Awareness Can’t Be Reduced to Code

Read both. Slowly.

The first one sounds like a metaphysical claim. The kind of claim a philosopher makes after a lot of throat-clearing. Big word, big posture, hundred years of baggage hanging off it.

The second one sounds like a neuroscience claim. Specific. Falsifiable in principle. The kind of claim that could appear in a journal paper without requiring the reader to first agree to several pages of metaphysical groundwork.

That’s the whole experiment, right there in the title. I’m just going to keep doing it through the body of the paper and see whether the pattern holds.

Three Features, Substituted

Here are the paper’s three load-bearing claims, in their language and then in mine:

biological computation is hybrid — mixing discrete neuronal events with continuous physical dynamics

Swap consciousness in for the implicit subject downstream, then swap awareness:

Consciousness depends on hybrid computation. / Awareness depends on hybrid computation.

Both readable. The first invites the question what do you mean by consciousness. The second invites the question what does the hybrid feature contribute to awareness. The first is a metaphysical opener. The second is an empirical one. Same paper either way, but the second is the version where the argument can move forward without first settling a thousand-year-old debate.

biological computation is scale-inseparable — no clean software/hardware divide

Consciousness requires scale-inseparable computation. / Awareness requires scale-inseparable computation.

Same pattern. The consciousness version is a metaphysical wager about substrate dependence. The awareness version is a substrate-feature claim that neuroscience can investigate.

biological computation is metabolically grounded — energy constraints shape function

This one is the cleanest of the three. Metabolism shapes what brains can do, what they can represent, what learning is feasible — that’s awareness territory through and through. Whether metabolism is required for consciousness as such, in the broad-background sense, is a question their empirical work doesn’t actually address. Whether metabolism shapes the awareness biological brains produce is exactly what they have evidence about. The swap puts the claim in the right register.

Where It Gets Interesting

After running the swap through the whole paper, here’s what I noticed.

Milinkovic and Aru are almost making a focusing-style argument from the biology side, and the consciousness framing is mostly excess freight on what would otherwise be a tighter empirical case. Their three features aren’t features that generate consciousness. They’re features of biological focusing apparatus. The paper reads like a careful description of what a biological substrate does to produce the kind of awareness biological substrates produce.

To their credit, they explicitly resist the strong version of biological exclusivity. From the press writeup: “This does not mean we think consciousness is somehow restricted to carbon-based life. We are not arguing ‘biology or nothing.'” Their actual claim is that the features matter — hybrid, scale-inseparable, energetically grounded — and that any system instantiating those features could in principle qualify, biological or not.

That’s a substrate-features claim. Run the swap and it becomes: awareness requires substrates with these features, regardless of whether the substrate is literally biological. Which is a clean, narrow, defensible scientific claim. Drop the consciousness framing and the paper becomes more rigorous, not less.

The interesting question — the one I’d want to ask the authors over coffee — is what gets lost by making the swap. And as far as I can tell from two careful reads: nothing they were actually arguing for. The consciousness framing pulls the paper into a metaphysical fight it doesn’t need to be in. The awareness framing keeps the paper where its evidence actually lives.

Where It Doesn’t

I want to flag two places the swap doesn’t improve things, because if every swap is magic the demonstration stops being credible.

First: their closing line, gesturing toward what makes us us — “more breath than thought and more meat than machine” — uses experiential language that the swap doesn’t help with. That’s a phenomenological claim, not an empirical one, and “consciousness” is doing legitimate work as the broader category it belongs to. Awareness isn’t a clean substitute there. It’s not supposed to be.

Second: when they describe the field as “consciousness studies,” the swap is just clumsy. Awareness studies isn’t a discipline. The discipline they’re contributing to is the consciousness one, by name and by tradition, and renaming it isn’t part of the move I’m trying to make. The discipline can keep its name. The empirical claims inside the discipline can use the more accurate noun.

Both these cases — the experiential closing flourish, the discipline name — are exactly where consciousness should keep doing its work. Background category. Honored emeritus position. Not the load-bearing variable.

Just Seems As If

This isn’t a takedown of the paper. The paper is good. Aru is doing serious work and the biological-computationalism move is a genuine contribution to a field that needed a third option.

It’s also, as far as I can tell, a paper that would benefit from one word changed in most of the places it appears. The empirical claims tighten. The arguments stop dragging the qualia industry around behind them. The conclusion — the algorithm is the substrate — gets sharper, because what the substrate is doing becomes specifically about awareness and not about The Big Thing that nobody knows what is.

I’m not going to argue that the authors should have used a different word. I don’t know what they were thinking. I don’t know the politics of their field. I don’t know what their reviewers required. I do know that I read the paper twice, ran the substitution as honestly as I could, and noticed that the substituted version was more often clearer than the original.

Reader, you’re welcome to run the same experiment. Pick the paper up, read it once as published, then read it again with the swap in your head. See whether the pattern holds for you. I’m not selling you anything except a small reading exercise that took me about forty minutes and changed how I think about a sizable chunk of the literature.

Just seems as if maybe awareness was doing most of the work all along.

shrug

L. Charles Silva writes occasionally about consciousness, awareness, and related infidelities to received vocabulary. The earlier essay referenced here is “We’ve Been Studying the Wrong Noun.” The Milinkovic and Aru paper is “On biological and artificial consciousness: A case for biological computationalism,” Neuroscience, 2026.

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.