ACT on the thought you keep arguing with

You have been arguing with yourself for twenty minutes, and the argument is going well. You have the evidence that you did fine in the meeting. You also have the evidence that you didn't, and you keep finding more of it, because you are the one looking.

Whichever side wins tonight, you will have spent the evening in court.

ACT starts from the observation that the court is the problem, and that it does not matter much who wins.

In one sentence

Acceptance and Commitment Therapy says the fight with a thought costs more than the thought. So it leaves the content of the thought alone and works on your relationship to it.

The problem it tries to solve

You have a difficult thought or feeling, so you try to get rid of it. That is the sane response, and with most problems it works: if something is wrong, you fix it.

Applied inward it misfires. Trying not to think about something requires checking whether you are thinking about it. Reassuring yourself that the fear is unfounded teaches you that the fear needs answering. Avoiding the situation that triggers it shrinks the part of your life you can still walk into.

ACT calls the pattern experiential avoidance, and its claim is not that avoidance is weak or irrational. Its authors describe it as what happens when the ordinary human talent for solving problems is turned inward onto experience, where it does not apply. Its claim is that this is expensive: the effort works in the short run and charges you in the long run, and the bill is usually paid in the things you stopped doing.

How it sees the mind

Two layers, and they are worth keeping apart, because only one of them is asked of you.

The model. ACT sits on a theory of language called Relational Frame Theory. It holds that humans learn to relate anything to anything, with the relation set by the context rather than by what the things themselves are like; that a relation learned in one direction entails one coming back; and that relations combine into networks nobody taught you. Then the claim that matters here: what an event does to a person can travel along those relations to the words tied to it. That, on this account, is why a thought can hurt in something like the way the event hurts. It is a research programme with its own experimental literature, and its adequacy is argued about by people qualified to argue about it.

The practice. Nothing in doing ACT requires you to believe any of that. The instruction is not "understand why words hurt", it is "notice what happens when you obey this one". That is not a concession: ACT's own stated test of an idea is whether acting on it works, not whether it is true, which is a strange thing for a theory to say about itself and the reason it can afford not to be believed.

Kept at the practical layer, the description is small. A thought is an event occurring in you, not a report you are obliged to answer. It may be true. It may be false. That is a separate question from whether obeying it is getting you anywhere, and ACT has a word for the second question: workability.

A worked example of the relations, since the abstract version is heavy: if you learn that A is bigger than B, you know without being told that B is smaller than A. Nobody taught you the return trip. Now make A a word and B an afternoon you would rather not remember, and you have the shape of the problem.

The central move

Stop working on the content and start working on the relationship.

ACT declares six processes, not one, and its practitioners will tell you the six are a single thing seen from six sides: acceptance, cognitive defusion, contact with the present moment, self as context, values, and committed action. Together they are called psychological flexibility, which its authors define as being able to stay in contact with the present moment and to change or persist in what you are doing when that serves something you have chosen.

The one the others hang from, for the problem this library is about, is cognitive defusion: seeing a thought as a thought rather than as the world it describes. The clumsy sentence that does most of the work is I'm having the thought that I did badly in the meeting. Same words, a frame around them, and the frame is the whole intervention.

Then the move that keeps defusion from becoming a trick: values. Dropping the argument is not the goal. Dropping the argument in order to do something you actually care about is. Without that second half, ACT is a relaxation technique with unusual vocabulary, and its own authors say the workability question cannot be asked without it: working has to mean working towards something.

A case where the move is hard

Not a case of a person. The place where this gets difficult is inside the literature, and it is better to say so than to invent a patient.

Is ACT actually doing what it says it does? ACT claims to work by increasing psychological flexibility, and that claim is testable: if the mechanism is real, the therapy should improve outcomes through that change and not through something else. Studies designed to test the mediating step exist, and most of them come out the way the theory predicts. The catch is in how they are built: the reviews that collected them say so themselves, and the recurring flaw is that the change in flexibility and the change in outcome are measured at the same moment, so nothing in the data shows which one moved first. That leaves a literature that agrees with the model and cannot yet confirm it, which is not the same thing as a literature that supports it.

Is it different from the cognitive therapy it defines itself against? This is the older argument, and it has a name: whether the acceptance-based approaches are a genuine new generation or a repackaging. Stefan Hofmann and Gordon Asmundson, two clinical psychologists writing from outside the school, put the question in the title of a 2008 paper, new wave or old hat? Sympathetic reviews have concluded that ACT does about as well as established cognitive behavioural therapy and no better. Its own authors said so first, in 2006, and the meta-analyses since have not moved it.

And is the evidence as strong as the enthusiasm? Lars-Göran Öst, a Swedish clinical psychologist who audits trials for a living, has reviewed this literature twice. In 2008 he looked at the whole third wave, the family of acceptance-based therapies that ACT belongs to, and found they used a methodology significantly less stringent than the CBT studies he matched them against. In 2014 he looked again at ACT alone and reported no improvement in methodological quality since, a smaller average effect than before, and no disorder for which ACT was yet well established.

What happened next is the interesting part. Ten authors, among them Steven Hayes, the psychologist who built ACT, published a line by line response documenting what they counted as ninety one errors across most of the studies reviewed, and asked for the review to be set aside. Öst answered in the same issue and withdrew nothing. The journal added an editor's note. Neither side moved.

Twelve years on, the most recent meta-analysis of ACT against other established therapies still lists the same shortcomings by name: trials not preregistered, too small to detect the difference they are testing, and read more favourably than their own numbers allow. So the question is not settled and it has not gone quiet, worth knowing before you read anybody's summary of the evidence, including this one.

None of this makes the practice useless, and nobody in the argument says it is. What is in dispute is the size of the claim, and whether the evidence yet earns it a place ahead of the alternatives. A tradition two and a half thousand years old cannot be wrong in this particular way, because it never promised a measurable effect. ACT did, and that cuts both ways: promising is what lets somebody check, and it is also what leaves you exposed when they do.

What you actually practice

Six processes, and the reason there are six rather than one is that they are meant to catch each other's failures.

Acceptance. Allowing a feeling to be present without spending effort on removing it. Not approval, not resignation: the specific act of stopping the struggle.

Cognitive defusion. Putting a frame around the thought. I'm having the thought that..., or saying the sentence slowly, or noticing the mind doing the thing it does at this hour.

Contact with the present moment. Attention to what is actually happening, rather than to the rehearsal of what happened or the preview of what might.

Self as context. Noticing that the one observing the thought is not the thought. This is the piece that sounds most like the contemplative traditions, and the resemblance is not accidental: the authors have written about the overlap themselves.

Values. Naming what you want your life to be about, in terms specific enough to act on today. Not goals: a goal can be completed, a value is a direction.

Committed action. Doing the thing, while the unpleasant internal event is still present. This is the one that makes the previous five something other than a way of feeling better in private.

What they aim at, in the tradition's own terms: not comfort, but the ability to keep moving in a chosen direction with the discomfort still there.

What people often misunderstand

Acceptance is not resignation, and it is not approval. The authors put it as a distinction between willingness and wanting: you may not want to do the thing you said you would do, and you can do it anyway. What you stop spending effort on is removing the internal event. It does not mean you accept the situation, stay in the job, or put up with how you are being treated. What ACT asks you to accept is defined narrowly, as the private stuff, and the outside world is exactly where it expects you to act.

Defusion is not a technique for making thoughts go away. The moment you use it that way it stops working: checking whether the thought has gone means you are still keeping score in the argument you meant to leave. Its authors describe the result as a decrease in how believable a thought is, or how attached to it you are, rather than an immediate change in how often it shows up.

It is not the power of positive thinking with new words. ACT does not ask you to replace a thought with a better one. It declines to arbitrate the content at all, which is a stranger move than optimism and is regularly mistaken for it.

"Workability" is not "whatever makes you feel good". The criterion is whether obeying the thought moves you towards what you have said matters, and that frequently recommends the uncomfortable option. Read as permission, it becomes the opposite of the therapy.

And the misreading specific to this one: because ACT is a clinical protocol, its language travels into places that are not therapy. This library included. See the last section, which is about a conflict of interest rather than a technique.

Where it sits next to what came before and after

Before: behaviour therapy, then cognitive therapy. ACT comes out of that lineage and then declines its central move. Classical cognitive therapy examines a thought for accuracy and tests it against evidence. ACT holds that the examination is itself part of the trap, and changes the relationship instead. Two therapies from the same family that disagree about whether the content of a thought is the business.

Beside: the contemplative traditions, and here the authors have been the ones drawing the line rather than denying it. Hayes published a paper on ACT and Buddhism in 2002. Self as context is recognisably a cousin of what Zen and Stoicism describe, arrived at from a laboratory rather than from a monastery or a school.

But the reasons stay incompatible, and that is why they are worth putting side by side. Stoicism wants the judgment to be true. Zen questions the self that is demanding a verdict. ACT sets truth aside as the wrong question and asks what obeying the thought is costing you. The instruction these traditions arrive at looks similar; the positions behind it cannot all be right, and the positions are the part you can actually use.

After: ACT is now one of the standard contextual behavioural therapies, delivered in health services with the ordinary apparatus of trials, manuals and training. In England the national guideline for chronic primary pain, as of 2021, names ACT and cognitive behavioural therapy in the same sentence, delivered by staff with appropriate training. That is the part that dates. Everything in this piece about what ACT says will read the same in ten years. What is known about how well it works will not, and this page carries a review date for that reason.

Go deeper

• Hayes, Strosahl and Wilson, Acceptance and Commitment Therapy (1999, subtitled An Experiential Approach to Behavior Change; second edition 2012, The Process and Practice of Mindful Change): the source text, written for clinicians.
• Hayes, Luoma, Bond, Masuda and Lillis, Acceptance and Commitment Therapy: Model, processes and outcomes (2006): the compact statement of the six processes, by the authors.
• Hofmann and Asmundson, Acceptance and mindfulness-based therapy: new wave or old hat? (Clinical Psychology Review, 2008, doi 10.1016/j.cpr.2007.09.003): the sympathetic but unconvinced case. Note the doi: the same author published a second, similarly titled piece the same year.
• A-Tjak et al., meta-analysis of ACT for mental and physical health problems (2015), and Gower, Styler and Litz, ACT against other bona fide psychotherapies (2026): the friendly summary and the unfriendly one, eleven years apart, arriving at the same answer for the comparison with cognitive behavioural therapy.
• Öst, on the methodology of the trials (2008, on the third wave as a whole; 2014, on ACT alone): the criticism from inside clinical psychology.
• Atkins et al., response to Öst (2017), with Öst's rebuttal in the same issue (2017): read them together or not at all. They are the reason the evidence question is not settled by counting studies.
• Association for Contextual Behavioral Science (contextualscience.org, consulted August 2026): the authors' own glossary, which is where the terms in this piece are defined.

What this changes in how you'd use Kitehold

One disclosure first, because it changes how you should read this section.

Kitehold is built on an idea from this tradition, and says so in its own product document. The founding metaphor, the kite, is there to carry one principle, and the spec names it: cognitive defusion, the thought stays and your relationship to it changes. The app's strongest internal rule, that a recorded thought is never deleted or made to look smaller than it was, is that same principle turned into a constraint. And the wording you meet in the app, I'm noticing the thought that..., is a defusion formula almost verbatim.

So this is not a neutral illustration of a tradition by a product that happens to fit it. On this page the tradition is describing the house it lives in. Discount it accordingly, and read the rest of this section knowing that I wrote it knowing you would.

With that said, the concrete thing:

The pull question is not the belief question. Kitehold asks how strong the pull is, and separately, elsewhere, how much you believe the thought. Two scales, deliberately different. ACT sets the truth of the thought aside and asks something else: what is obeying it costing you, measured against what you have said matters. The pull number is not that question answered, and it should not be read as one; it measures the push to obey, not the case for obeying. A thought can be completely accurate and still be running the night, and if you have ever been right and miserable at the same time, you have already had the experience the distinction is pointing at.

And the thing the number cannot tell you is the one ACT cares about most: whether the evening went where you had said you wanted it to go. A pull that drops is not the same as a night spent on what matters, and the app has no field for the second one. That part is yours to notice, and then to go and do the thing you would have been doing anyway.