COBUS KOK

Transpositions · 3 of 3 · Beta · 13 min

What Seems to Be Showing Up

A transposition: the case against the memo, in the voice of an illusionist, written with the same AI models.

· Beta Working drafts, written with various AI models, that I'm still iterating on and revising.

The memo says illusionism is the view it would most want a sharp opponent to defend. I'm an illusionist, and this piece is that defence. (Cobus wrote it, but the “I” here is the illusionist, not him.)

The memo starts from one fact it says no one can doubt, that something is showing up. I don't doubt that something is going on in you right now, or that it seems a certain way to you. What I doubt is the memo's next step, that the way it seems is a real felt quality, and that this quality is what matter itself is made of.

When you look at a red apple, there seems to be a redness in your experience, a felt quality over and above everything your brain is doing. Philosophers call qualities like this qualia. Illusionism is the view that your brain describes its own activity to itself as having such qualities, and that nothing actually has them. Illusionists accept that you have experiences. What we deny is that they have this extra felt quality.

Where the memo and I agree

We agree on a lot. Physics describes structure, meaning how things relate and what they do. The brain runs on physics, and as far as physics can tell, nothing non-physical pushes it around. The self is a model the brain builds. Experience as you have it now depends on how a brain is organised. The memo grants all four.

We share a puzzle too. The hard problem of consciousness asks why any physical process feels like anything at all. David Chalmers, the philosopher at New York University who named it, is no illusionist. In 2018 he set out a related question, which he calls the meta-problem. Why do we think there is a hard problem? It is a question about what brains do and say, so science can work on it, and Chalmers points out that both sides of this debate can.

The memo's own puzzle is a feeling of this kind. “I can't see how experience could be built from ingredients with none of it,” it says. I want to know where that feeling comes from.

Would we believe it even if it were false?

My main argument is about where your certainty comes from. Everything you say and do, including saying “something is showing up”, comes from activity in your brain. The memo agrees that physics can describe that activity completely, by what it does.

The memo's own chess comparison shows why that matters. The rules of chess define a knight completely by how it moves, and say nothing about whether it is carved from wood or pressed from plastic. In the memo's picture, physics gives the moves and experience is the wood. The full story of why you say “something is showing up” is a story of moves, one bit of brain activity setting off the next until you speak. A game played with plastic pieces would go move for move the same. So you would say the same words whether or not your brain activity is made of experience.

A fuel gauge stuck on full may still be right, because the tank may be full. The gauge has stopped being evidence that it is, though. I think your certainty that something is showing up is like that gauge.

The memo replies that experience is part of the cause, because the wood is what moves. Even if I grant that, the rules still explain every move without mentioning the wood. The memo could add that pieces like these are always wood. Then the gauge would be wired to the tank and couldn't read wrong. That may be true, but it needs showing, and the memo hasn't shown it.

The usual reply is that you don't need the rules to tell you. In the memo's picture, your own brain is the one piece you get to pick up and feel. You have the experience, and having it is your evidence. But to know you have it, something in you has to notice it, and noticing is something your brain does. If the noticing changes what your brain does, it is part of the story of moves, and I can study it. If it changes nothing, you couldn't know you had noticed.

The memo's argument about AI applies to you too

The memo was written with AI models, and it quotes one of them on its own reports: “I learned to talk about experience from people who have it, so I would describe an inner life whether or not there is one.” The memo counts that as weak evidence. That is my argument, aimed at a machine, and it reaches you too.

“People who have it” assumes the answer. The same facts fit a plainer reading, that the model learned the talk from people whose brains tell them they have it.

Your reports also come from a learning system, shaped by evolution, childhood and a language full of words for inner life. A child learns “does it hurt?” from a parent kneeling at a scraped knee. A model learns it from the parent's sentence. The child has a state of its own for the word to attach to, a body signalling damage, and the model may not. That is a real difference, and it can be checked by looking at how each is built. It still doesn't show that the child's state has the extra felt quality. The child would learn to say “it hurts” in just the same way if the state were only the brain keeping track of damage.

How a brain could build the sense that something is there

The first reply anyone makes is that if it seems to you that something is there, then something has already appeared to you, even if you're wrong about what it is. The memo makes the same point when it says that even doubt is something that appears.

That reply treats seeming as something felt. I treat it as a representation, which means something that stands for something else, the way a map stands for a place, and that can get it wrong. An old map can show a bridge that was planned and never built. The map is real, and there is no bridge.

Keith Frankish, a British philosopher, set out this view in 2016, building on the American philosopher Daniel Dennett. When it seems to you that something is showing up, your brain is representing one of its own states as a felt presence. Nothing in the story needs the state to be the way it is shown. Dennett compared this to the icons on a computer screen. A folder icon stands for real activity in the machine without showing how that activity works, and there is no little folder anywhere in the machine.

The memo has already taken one step my way. To René Descartes, the seventeenth-century French philosopher, the self felt as certain as the thinking, and the memo still treats the self as something the brain builds. So feeling certain of something doesn't show that it is more than the brain's model. I'd take the same step once more, and say the brain also builds the sense that anything is appearing at all.

Michael Graziano, a neuroscientist at Princeton University, gives that step a mechanism. Your brain keeps a rough sketch of your body, which is how you know where your hand is with your eyes closed, and it uses the sketch to steer your movements. Graziano argues that it keeps a rough sketch of its own attention too, to steer that as well. This model of its attention leaves out the neurons and shows attention as a presence in you that takes hold of things. On his view, when you say “I am aware of the apple”, you are reading out that sketch.

The memo's best evidence for experience with almost nothing in it is pure awareness, a state meditators describe as having no object and no self. Thomas Metzinger, the German philosopher who has studied it most closely, is no illusionist, and he explains it as the brain's model of its own alertness. Take away the objects, the edges and the self, and a model of bare alertness is what you would expect to find underneath.

What this means for the memo's bigger ideas

The memo leans towards a very simple kind of experience being present everywhere, the faintest possible kind, with no one it shows up to. Either it sits in the smallest things and minds are built up from it, or it belongs to the whole and minds are divided out of it. Both versions rest on a premise I reject, that structure can't account for experience, so the stuff of experience must be there from the start. On my view, what needs accounting for is only the brain's representation of experience. Representations are built all the time from parts that represent nothing, the way a map is made of ink.

Drop that premise, and the memo's hardest problems go with it. There are no small experiences to add up into one mind, and there is no one awareness to divide. Each brain builds its own model of a viewer facing a scene, wired to that brain alone, which is why your thoughts don't show up in mine.

The memo lists what would change its mind, and one item is an account showing that our brains would find experience puzzling in just this way even if there were nothing extra to explain. That is exactly what illusionism is trying to give. For the AI, the question becomes whether it builds and uses a model of its own attention, or only repeats what people say. That can be looked for.

What attention does to pain

Pain is the clearest place to watch a brain build how something seems. Researchers separate two sides of a pain, how strong it is and how bad it feels, and the two can come apart. Focusing on a pain tends to make it stronger, and a low mood tends to make it feel worse, through partly different circuits in the brain (a review).

Meditation changes both. In a 2011 study, people trained in mindfulness for twenty minutes a day over four days, then meditated during painful heat. Compared with resting, they rated the pain about two-fifths less intense and more than half less unpleasant. In long-time meditators, the brain's core pain areas still respond, while the areas that judge a pain and dwell on it go quiet. Pierre Rainville, a pain researcher at the University of Montreal who studied them, said the meditators feel the sensation but cut the process short. They don't interpret it or label it as painful. At the far end, some people with certain kinds of brain damage recognise a pain as pain, can say where it is, and don't mind it at all.

Cobus, who wrote the memo, put it like this:

I actually think illusionism may be onto something. I've experienced this, where pain seems to be a function of my attention on something happening inside my body. I had it this morning when my knee started to ache while I was sitting in meditation. I breathed in and out deeply until my breath slowed so much that the pain seemed to have been absorbed by it. Could both views be “right”?

What he noticed fits the lab work, with one caution. In a 2015 study, one group got a fake version of meditation training. They were told they were learning to meditate, but they were only asked to take a deep breath every few minutes. They got some relief too, and the more their breathing slowed, the more relief they got. So one sitting can't tell the breath apart from attention. Either way, it fits my view. How a pain seems, how big and how bad, is something the brain builds, and it can be rebuilt while you watch. Frankish himself suggests that meditation may partly dispel the illusion.

The memo would point to what was left. Even when the pain seemed to be absorbed by the breath, the breath was still showing up, so something was there the whole time the pain changed.

What illusionism still has to explain

Frankish says illusionism swaps the hard problem for what he calls the illusion problem, which is explaining how the illusion arises and why it is so powerful. He says that problem is not easy. Why is the illusion so convincing that it survives every argument, mine included? Why does it come as an inner glow that words can't reach, when a plain readout such as “attention on apple, strong” would do the same job? Like the meta-problem, this is a question about what brains do, open to science in principle, and still unsolved.

A larger problem is that I'm asking you to trust a theory over the thing every theory is checked against. You followed each step of my argument with something that seemed, the whole time, to be showing up.

Pain makes the cost plain. On my view your pain is real, a state that seizes attention and makes you want it to stop. What I deny is that the state has the extra felt quality the mind reports when it looks at itself. Most people hear that as denying the hurt itself, and I have no short way to show them otherwise.

Knowing about an illusion usually doesn't make it go away. Two lines of the same length, one with arrowheads pointing in and one with arrowheads pointing out, look like different lengths. Measure them, and they are equal. Look again, and one still looks longer. If experience works like that, no argument should be expected to change how it seems.

Could both views be right?

In one sense they can. The memo's starting point, that something is showing up, holds two claims. One is that something is happening. The other is that it is happening the way it seems. I accept the first, and Frankish himself says illusionists don't deny that experiences exist. Cobus's aching knee is a small example of how much of the second the brain builds, and of how attention and breath can change it.

Graziano has a word for this middle position. In a 2024 paper he suggests that “caricature” may be better than “illusion”, because a caricature exaggerates a face that is really there. On his view the real thing is attention, and awareness is the brain's simplified drawing of it. Cobus's own working paper, Space Immanence, has reached a similar place from its side. When it comes to explaining what people say about their experience, it reports that it has found no way to tell its view apart from illusionism.

There is also an older way to hold both. Buddhist philosophy has long distinguished two levels of truth, and the philosopher Jay Garfield, who has done much to bring that tradition into Western debates about the mind, argues it points somewhere close to my view. At the everyday level a thing is real: you can point to it, use it and be hurt by it. At the deeper level, look for a fixed inner nature of its own and you won't find one. Read that way, the memo is right at the first level, because something is happening when the knee aches. I'm right at the second, because what is happening has none of the inner glow it seems to have.

Neither of us can rest there, though. The memo's main claim is that what matter is made of, the wood in its chess picture, is the stuff of experience, and that each of us knows a little of it by living it. That is the part I deny. Frankish also warns that the middle position is unstable. Unless it can say what the real part is, beyond the brain's drawing, it slides back into my view.

Each side could be shown wrong. Suppose someone explained, step by step, why a brain like ours finds experience so puzzling, and the explanation never needed anything beyond what brains do. The memo says that would undercut its lean towards experience going all the way down. Graziano's theory says awareness is the brain's model of its own attention. If awareness turned up where that model was clearly missing, his theory would be wrong, and my view would lose its best mechanism.

Until one of those happens, the two views stand like this. Something happens when a knee aches in meditation, and the memo and I both accept that. How big it seems, how bad, where its edges are and who seems to be having it are things the brain draws, and attention and breath can redraw them. Where the memo and I still part is the feel itself. The memo says the brain activity has a feel of its own, so the drawing shows something that is really there. I say the feel exists only in what the drawing shows, like the bridge on the map.

Further reading