Someone tells you to relax and you become fractionally more tense. Someone tells you to stop worrying and the worry gains weight. You decide tonight you will definitely sleep, and you are awake at two.
None of this is a personal failing. These states share a structure, and it makes them immune to instruction.
The states that effort destroys
Most things respond to trying harder. Lifting, running, memorising, finishing a task — effort improves them all.
A small set behaves in the opposite way. Falling asleep. Relaxing. Forgetting something. Being spontaneous. Finding a word on the tip of your tongue. Each of these arrives only when you stop reaching for it, and each recedes in proportion to how hard you pursue it.
The reason is that effort has a physiological signature. Trying involves monitoring, evaluating, and a low current of urgency — a mild state of activation. When the goal is deactivation, the means and the end contradict each other. You are pressing an accelerator to slow down.
Relaxation is not something you do. It is what remains once you stop doing everything else.
Why telling yourself not to think about it fails
The psychologist Daniel Wegner described the mechanism as ironic process theory, and the classic demonstration is simple: try not to think of a white bear.
To follow that instruction, some part of you must hold a model of the forbidden thought so it can check whether it is present. That monitoring keeps the thought active. And under load — when you are tired, stressed, or under time pressure — the monitor outlasts the suppression, so the unwanted thought comes back more often than if you had never tried.
This is why "don't think about the meeting" reliably produces the meeting, and why it works worst exactly when you need it most: at 3 a.m., exhausted, with nothing else to occupy attention.
Three ways the effort trap shows up
Checking whether it's working. You do the breathing, and after ninety seconds you ask yourself whether you feel calmer yet. That question is an evaluation, evaluation is a task, and the task keeps the system on duty. It is the same loop as looking at the clock at 3 a.m.: the measurement sustains the problem.
Relaxation as a performance. When there is a right way to do it, there is a way to fail. People report tensing up during relaxation exercises for exactly this reason — the technique has become a test with a standard to meet.
Being annoyed at being tense. A second layer on top of the first. Now you are activated, and also frustrated about being activated, and the frustration is itself activating. Most people spend more energy on this layer than on the original stress.
The shift that works: aim at something you can actually do
The way out is not to try harder or to try less. It is to redirect the effort onto a target that accepts it.
You cannot decide to be calm. You can absolutely decide to make the next out-breath longer than the one before. That instruction is concrete, verifiable, and achievable on the first attempt — and the calm follows it about two minutes later, without ever having been requested.
The same substitution works everywhere: not "relax your shoulders" but "let them fall on the next exhale". Not "clear your mind" but "count ten breaths". Not "stop worrying" but "I'll think about this at six o'clock". Each replaces an unachievable state with an achievable action, and the state arrives as a by-product.
Permissive language, and why it isn't just softer wording
This is the practical core of the Ericksonian approach, and it is more than politeness.
Compare two instructions. Your arms are heavy. If they aren't, you have just been given a claim to disprove, and the disproving is alert, critical work. Versus: You may notice one arm feels a little heavier than the other, and you don't have to decide which.
The second cannot be failed. "May notice" makes it optional. "A little" sets a threshold you can actually meet. "Don't have to decide" removes the task. There is nothing to check, so the checking stops — and with it the low-level arousal that checking generates.
Three phrasings that do the same work: whenever you're ready, as much or as little as you like, you might find that…. If direct suggestion has always felt to you like being told a comfortable lie, this is why, and this is the version to use.
Give it something to do instead
Suppression fails; substitution works. An anxious mind with nothing to hold generates its own material, so the answer is not an empty mind but an occupied one.
Deliberately dull objects work best: the weight of your legs, the count of a breath, warmth in the hands, the pressure of your feet on the floor. None are interesting, and that is precisely the point — nothing competes, but attention still has somewhere to be. "Clear your mind" never provides that; it asks for an empty stage and then wonders why the worry walks onto it.
And when a thought arrives anyway, returning to the count is the practice, not evidence of failing at it. That single reframe removes most of the self-criticism people bring to this work.
Where this stops being enough
Everything here concerns ordinary stress and ordinary sleeplessness, where the effort trap is genuinely the main obstacle.
It is not the explanation for everything. If anxiety is shaping what you avoid, if panic attacks recur, if you have slept badly most nights for three months, or if low mood has come with it, then the problem is not that you are trying too hard to relax. Those have specific and effective treatments, and reaching for a subtler technique is a way of postponing them.
Keep reading
How to calm down fast — the principle applied: aim at the exhale, not at the calm.
How to fall asleep when your mind won't stop — the same trap at midnight, where it does the most damage.
Common questions
Why do I get more tense during relaxation exercises?
Usually because the exercise has become a performance with a standard. You are monitoring for results, and monitoring is a mild state of alertness. Switching to permissive wording — you may notice rather than you will feel — removes the standard and usually removes the tension with it.
Is thought suppression always harmful?
Not always, but it is unreliable and tends to fail under exactly the conditions where you need it: tiredness, stress, time pressure. Postponing a thought to a specific time works considerably better than banning it, because it gives the monitor something to accept.
Does this mean I should stop trying altogether?
No — redirect the effort rather than abandoning it. Trying to be calm fails; trying to lengthen an exhale, or to complete ten slow breaths, succeeds. The effort is useful as long as its target is an action rather than a state.
Is permissive language just weaker suggestion?
It is weaker as a command and stronger in effect for many people, because nothing about it can be contradicted by experience. Direct suggestion suits people who accept instructions comfortably; indirect suits those who instinctively test them. Neither is superior — they suit different temperaments.