There’s a particular misery in knowing better and doing worse.

You understand the situation completely. You could explain to someone else exactly why the thing you’re about to do is a mistake, and you’d be persuasive. And then you do it anyway, and afterwards the understanding is still there, entirely intact, having changed nothing.

That’s not weakness and it isn’t hypocrisy. It’s a structural feature of how a person is built, and once you can see the structure the whole thing becomes tractable.

You have three layers, and they can disagree with each other.

The Model

A person’s conduct is produced by three layers operating at different depths. The action layer is what is done. The mind layer is what is believed, reasoned and understood. The heart layer is what is wanted, valued and identified with. Causation runs predominantly downward — what a person wants shapes what they find convincing, which shapes what they do — while feedback runs slowly upward, since repeated action under conditions of genuine choice gradually alters what is wanted. Change applied at the action layer alone requires continuous effort and ends when the effort is disrupted. Change at the mind layer generalises more widely but fails wherever belief conflicts with the heart layer, since in a direct conflict the deeper layer prevails. Durable change therefore requires movement at the heart layer, which cannot be produced by decision and is instead produced by accumulated evidence from action.

Three layers, and the interesting part is what happens when they conflict.

The Layers

Action. What you do. Observable, measurable, and the only layer anyone else can see directly.

Mind. What you believe, know, and can argue for. This is where your understanding of the problem lives — including your entirely accurate understanding of why the thing you keep doing is a mistake.

Heart. What you want, what you value, what you take yourself to be. Not emotions exactly. Closer to appetite plus identity: the thing that determines what looks attractive to you before any reasoning starts.

Causation runs downward. What you want shapes what arguments you find convincing, which shapes what you do.

That last point is worth pausing on, because most people have the mind layer’s role backwards. Your reasoning is not neutral. It works for the heart layer, and it’s very good at its job. Which is why you can construct an excellent case for the thing you were going to do anyway, and why the case will genuinely convince you.

Three Conflicts

Here’s where the model earns its keep. The layers can disagree, and each disagreement has a characteristic feel and a characteristic resolution.

Action contradicts mind: rationalisation

You do something you believe is wrong.

This is uncomfortable — it’s the sensation of being a hypocrite — and it resolves. Not usually by the behaviour changing. By the belief changing to accommodate the behaviour.

You’ll notice this in yourself if you look. The thing you did that you thought was a mistake becomes, over a few weeks, a thing that was defensible in the circumstances, and then a thing that was probably right.

That’s not dishonesty. It’s a conflict resolving in the direction of least resistance, and belief is much cheaper to move than behaviour.

Practical implication, and it cuts both ways: doing something repeatedly changes what you believe about it. Which is useful when you’re deliberately building evidence for a new self-concept, and dangerous when you’re drifting, because the drift comes with a matching set of convictions that arrive on schedule.

Mind contradicts heart: knowing better, doing worse

The most common and the most miserable.

You believe you should. You don’t want to. The wanting wins.

This is the specific condition described at the top of this article, and it explains why insight so reliably fails to produce change. Your understanding is at the mind layer. The thing generating your behaviour is deeper, and in a straight conflict the deeper layer wins — not because it’s stronger in some mystical sense, but because it runs first and your reasoning arrives afterwards to explain the result.

This is why “I know what I need to do” is such a poor predictor of doing it, and why the most self-aware person in a room frequently isn’t the one whose life is working. The full version of that failure is in you already know, that changed nothing.

The tell for this conflict: you can argue your own case perfectly and it makes no difference. If understanding were the constraint, understanding would have fixed it.

Heart moves and action lags: being ahead of yourself

The third, and it’s the rare good one.

Something shifts in what you want, and your behaviour hasn’t caught up. There’s a period where you’re operating on old defaults while wanting something different, and it feels like impatience with yourself.

This is the direction change is supposed to run, and the lag is normal. Behaviour follows a heart-layer shift with a delay, because the old defaults are still installed and defaults don’t uninstall on their own.

If this is your conflict, the work is straightforward: supply the behaviour deliberately until it becomes default. That’s much easier than the other two, and it’s the state people are in when change suddenly starts feeling possible after years of it not.

What Each Level of Change Actually Buys

Now the practical claim.

Action-layer change is maintenance. You install a behaviour on top of an unchanged person. It works while you’re expending effort and stops when the effort is interrupted — travel, illness, a bad fortnight. Nothing underneath is generating it. That’s the collapse pattern in why you go back to who you were.

Mind-layer change is willpower. More durable, because a genuine belief change generalises across situations rather than needing to be applied case by case. But it’s still an override, and overrides fail at exactly the moments they’re most needed: when you’re tired, under pressure, or in a hurry, which is precisely when reasoning is least available.

Heart-layer change is free. When what you want has changed, the behaviour requires nothing. There’s no override, no maintenance, no negotiation. It’s just what you do.

Which produces the summary worth keeping:

Behaviour change without belief change is maintenance. Belief change without heart change is willpower. Only heart change costs nothing to keep.

How the Deepest Layer Actually Moves

The obvious objection: if causation runs downward, and you can’t decide to want something different, how does anything ever change?

Through the upward feedback loop, which is slow and real.

Repeated action, taken under conditions where you genuinely had a choice, gradually alters what you want.

Note the qualifier. Action taken under compulsion, or purely for an audience, or as a one-off, does very little. What moves the heart layer is a pattern of choices made freely, accumulated over time, which functions as evidence about what kind of person you are — and the evidence eventually revises the identity that generates the choosing.

This is why action still matters enormously even though it’s the shallowest layer. It isn’t only the output. It’s the input to the layer above it, on a long delay.

Which resolves the apparent paradox: you can’t decide to want something, but you can do the thing repeatedly and freely, and wanting follows. It takes considerably longer than anyone would like, and it’s invisible while it’s happening, which is the accumulation threshold.

The oldest version of this argument is four hundred years old. A man broke a prediction that had been accurate for twenty years, and what he did wasn’t install behaviours or adopt beliefs. He audited what he wanted, in writing, and then acted against it consistently for a decade until the wanting changed. That’s the method, and it’s this model with older vocabulary.

What This Is Not

It is not an argument that action doesn’t matter. It matters twice — as the thing that produces results now, and as the only available input to the deepest layer. The claim is that action alone, without the upward propagation, is maintenance rather than change.

It is not a claim that understanding is useless. Mind-layer work tells you what to aim at, and aiming at the wrong thing is the most expensive error available. It’s necessary and it isn’t sufficient.

It is not a three-step programme. You don’t work the layers in order. They interact continuously, and the useful thing is knowing which conflict you’re in, because each one has a different resolution.

The Cost

Two prices, and the first is the one people find hardest.

You have to look at what you actually want, which is frequently not what you’d want to want.

Mind-layer work is comfortable. You examine your beliefs, find some errors, correct them — and your self-image survives, because being wrong about a belief is respectable.

Heart-layer work means naming appetites. Needing to be seen as the capable one. Wanting the argument more than the resolution. Preferring the comfortable version of your life to the one you claim to be pursuing.

Nobody enjoys writing those down, and writing them down is the step.

Second: the timeline is measured in months and produces nothing visible for most of them.

The upward loop is slow. There’s no week where you can tell it’s working, and every fortnight you’ll have the option of returning to action-layer fixes, which produce immediate visible results and don’t hold.

Most people take that option, repeatedly, for years.

Common Questions

Why doesn’t behaviour change stick?
Because behaviour installed on top of an unchanged person requires continuous effort, and any interruption ends it. Nothing underneath is generating the behaviour, so when the effort stops there’s nothing left.

Why do I know better and still do worse?
Because understanding sits at the mind layer and behaviour is generated deeper. In a direct conflict between what you believe and what you want, the wanting wins — it runs first, and your reasoning arrives afterwards to explain the outcome.

What’s the difference between changing behaviour and changing belief?
Behaviour change is maintenance and ends when effort is disrupted. Belief change generalises across situations but still functions as an override, and overrides fail when you’re tired or under pressure — exactly when they’re needed.

Which level of change actually lasts?
The deepest. When what you want has changed, the behaviour requires no maintenance because there’s nothing to override. That’s the only level at which change is free to keep.

If I can’t decide what to want, how does anything change?
Through the upward loop. Repeated action taken under conditions of genuine choice accumulates as evidence about what kind of person you are, and eventually revises what you want. It’s slow and invisible while it’s happening.

Why does doing something repeatedly change what I believe about it?
Because a conflict between action and belief resolves in the direction of least resistance, and belief is cheaper to move than behaviour. This is useful when building evidence deliberately and dangerous when drifting, since the drift arrives with matching convictions.

Where to Start

Ten minutes, and it’s a diagnosis rather than a plan.

Take the thing you keep not changing and work out which conflict you’re in.

Can you argue your own case perfectly and it makes no difference? That’s mind contradicting heart. You don’t have an understanding problem. Stop reading about it.

Have your beliefs about it quietly softened over the last year? That’s action contradicting mind, resolving downward. Worth noticing, because the softening is the mechanism rather than a change of view.

Do you want something different and keep defaulting to the old behaviour? That’s the good conflict. Supply the behaviour deliberately; the wanting is already there.

Then, if you landed on the second conflict — which most people will — one further question, written down:

What do I actually want here, that makes the current behaviour make sense?

Not what you should want. What would have to be true about your appetites for your conduct to be the obvious choice.

Write the plain version. It’ll be shorter and less flattering than you’d like, and it’s the only thing on this page that’s load-bearing.

I’ve spent twenty-five years working on how ordinary people develop capabilities that look impossible from the outside, and the ones who change permanently almost never describe having finally understood something.

They describe, usually with some puzzlement, that at some point they stopped wanting the old thing.