Nobody is in charge in there — many sub-minds run in parallel and post to one shared board. Once you see that, every technique in this book turns out to be the same move: put an intention on the board, often enough that the system adopts it.
Thirty seconds ago you decided — clearly, deliberately, with something that felt very much like a decision — to watch the breath. You are now three steps into planning dinner, and you have no memory whatsoever of agreeing to that.
This chapter is the theory that makes the rest of the map work. It's the model Culadasa lays out in his Fifth Interlude, and it sits here, before Stage 6, because the next two chapters lean on it entirely: why you ignore distractions rather than suppressing them, and why effortlessness is something the system does rather than something you achieve.
It also happens to be the most useful thing in the book for anybody who has ever debugged a distributed system.
Start with the raw evidence, before any model. You set an intention: watch the breath. A little while later, attention is somewhere else entirely — and the striking part isn't that it moved. It's that you were not consulted. There was no moment of choosing dinner over the breath. There was no felt override. The switch just happened, and then a version of you turned up afterwards to find it already in progress.
Then look at the return. You come back to the breath — and that didn't feel like a decision either. The noticing arrived. Something woke up and flagged it, and only after the flag went up did anything resembling "you" get involved.
Now try the question that model-building actually turns on. If there were a single boss in there, one decider running the show, how could this happen? A boss who decided to watch the breath would watch the breath. A boss can't lose an argument with itself, forget its own instruction, or be surprised by what it's doing. The everyday experience of sitting down to meditate is, on its own, fairly strong evidence that nobody is at the wheel — or rather, that a great many hands are, and they don't agree.
That's the fact. Here's the model that accounts for it.
Picture the mind not as one processor with a single thread of control, but as a large number of specialised parts, each doing its own job at the same time as all the others. Culadasa calls them sub-minds, and the whole picture the mind-system model.
Some of them are sensory sub-minds — one broad department per sense — and their work is enormous and completely invisible to you. Photons hit a retina; what arrives in your experience is "a car went past." Every bit of that transformation happened somewhere you have no access to. There's a discriminating (thinking) mind made of many sub-minds of its own: comparing, planning, judging, remembering, worrying. Among them is the narrating sub-mind, which stitches everything into a story with a main character in it, and which you have spent your whole life mistaking for yourself. And underneath all of it runs a great deal of deep unconscious processing that never surfaces at all — until, on a quiet enough cushion, some of it does.
For a programmer the shape is familiar. This isn't a monolith with a main() that calls everything else in order. It's a fleet of services: specialised, concurrently running, each with its own local state and its own priorities, none of them individually conscious, all of them doing their work outside consciousness. Consciousness is not where the computation happens. It's a much smaller thing, and section 3 is about what it actually is.
One guardrail before you go further: sub-minds are not little people. Nobody in there is scheming. They're processes with jobs — and when one of them hijacks your beam with a dinner plan, it is not being disobedient. It's doing exactly what it exists to do, at a moment you'd rather it didn't. Click around the diagram to meet them.
Here's the part that reorganises everything. Sub-minds don't hold references to each other. There's no call graph, no sub-mind reaching into another one to read its state. What they have instead is a common space: each one posts its output there, and each one reads what others have posted.
That common space is consciousness. It's a blackboard — or, if you prefer the newer vocabulary, a pub/sub bus with exactly one topic and a very small queue. "You experience X" means, in this model, something quite specific and slightly deflationary: some sub-mind's output got published to the board, where every other subscriber can now see it.
Notice what consciousness therefore doesn't do. It doesn't decide. It doesn't compute. It doesn't perceive — the sensory sub-minds already did that, and what shows up on the board is their finished report. Consciousness is an interface: the one place where information becomes available to the whole system at once, rather than staying locked inside the service that produced it.
Which is exactly why it's so expensive. One shared channel, limited bandwidth, and everybody wants time on it. That constraint is the same one you met in Chapter 2 as a single budget of conscious power split between the beam and the field: narrow the beam by hogging the channel and the field goes dark, because there's only ever one board.
And it's why broadcasting matters so much for training. Information that reaches the board reaches sub-minds that would otherwise never hear from each other — the part that wants to meditate and the part that keeps rehearsing yesterday's argument. Nothing gets learned across the system unless it goes through consciousness first. This is the whole reason conscious practice changes anything.
Sit for a moment and watch the instant just before attention moves. There's a small lean in it — a tilt toward the next thing, arriving slightly ahead of the movement itself. It isn't a decision, and it doesn't feel like one. It's more like a bid that has already been accepted by the time you notice it.
In this model every moment of consciousness carries an intention (cetanā), including the intention to keep doing nothing in particular. Sub-minds post theirs; the posts compete; one of them gets executed, and that's what the next mind-moment is. Attention moved to dinner because a planning sub-mind's intention won the slot. It's not that you failed to hold on — it's that a different message got picked up off the queue.
This makes intentions the only API you have. You can't reach into a sub-mind and set its state. There's no privileged console, no root access, no chairman's veto — the "you" that would issue the order is itself just another participant. What you can do is post an intention to the board, where every sub-mind in the system will read it.
So the usual framing of willpower is a category error. Straining harder isn't a stronger command; it's the same message shouted, which mostly just adds a second message about how badly it's going. What works is being persistent and consistent rather than forceful — because of what happens next.
Post the same intention to the board once and very little happens. Post it a thousand times, gently, across a hundred sits, and something quietly changes: sub-minds that kept reading that message start to adopt it as their own. And a sub-mind that has adopted an intention executes it without needing you — without any conscious effort at all.
You know this shape already. A hot path gets interpreted, then profiled, then compiled; afterwards the same work happens at a fraction of the cost and nothing in the foreground has to think about it. Or: you stop scheduling the job by hand because a daemon picked it up. Skills compile. What was expensive and conscious becomes cheap and automatic, and the capacity it was eating gets handed back.
This is the entire mechanism behind the ten stages. Every antidote in this book, at every stage, has the same underlying form — form this intention, repeatedly and gently, and let the system take it over. Intend to notice when attention wanders (Stage 2–3). Intend to catch dullness while it's still subtle (Stage 5). Intend to stop feeding the subtle distractions (Stage 6). You are not building willpower; you are running a training loop, and the reps are what count.
Gentleness isn't decoration here, it's load-bearing. Harshness posts a second message alongside the first — this is going badly, you are bad at this — and sub-minds adopt that one just as faithfully. Which is also why Chapter 3's odd instruction to enjoy the moment of waking up turns out to be mechanism rather than niceness: pleasant outcomes are what make an intention worth adopting.
And it sets up Stage 7 exactly. Effortlessness isn't a state you finally strain your way into. It's what a sit looks like once enough of the committee has adopted the intention that no conscious effort is left to spend. Drag the slider and watch the handoff.
Now put the pieces together. When sub-minds are pulling in different directions — one wants the breath, one wants dinner, one is defending you in an argument from Tuesday — the shared board is a shouting match. Conscious power gets spent on the friction between competing intentions, the beam gets hijacked several times a minute, and the field is too busy refereeing to stay bright.
As more of the committee adopts the same intention, that friction drains away. The technical name is unification of mind, and it's the second great axis of the training, running alongside mindfulness the whole way up. Same system, same total power, pointed one way instead of five. That's what makes the beam unhijackable in Stage 6, and what makes it hold itself in Stage 7.
A committee where everybody talks over everybody, versus a choir on one note. The choir isn't quieter because voices were removed — it's louder, and it's one sound. Increasing unification together with increasing mindfulness is precisely what ripens into the five threads of śamatha from Chapter 1: stability, then brightness, then joy, tranquility, equanimity.
Two cautions worth carrying forward. Unification is not suppression: a sub-mind you shout down hasn't stopped posting, it has just stopped posting where you can see it — which is the whole reason Chapter 8's instruction is ignore, don't suppress. And a mind that grows quiet enough starts giving deep unconscious material the bandwidth to surface. Old grief, old anger, things you'd have said were dealt with. That isn't a malfunction; that's the board finally being free enough for a long-queued message to get through.
This one is theory, but it changes what you do on the cushion more than most of the practical chapters do. Three cards: how the model shows up in an actual sit, how to meet the committee without turning it into a menagerie, and how to work the board rather than fight it. The leading word on each line is the cue; the note after it tells you what to feel for.