Letting go of a grudge: the body releases before the mind does

Letting go of a grudge does not start with a decision. Nobody stops resenting someone because they explained to themselves that resentment is bad for their health. A grudge holds on because it is doing something, and it only gets put down once that something is covered another way.

This is what most advice on the subject leaves out. It assumes a switch. Write a letter, burn the paper, picture a balloon floating off: none of that is absurd, but it addresses the part of you that already agrees. The part holding the grudge does not agree at all. It has a reason, and it will not trade that reason for a metaphor.

There is a second misunderstanding, and it is stubborner still. People assume that letting go of a grudge means forgiving, and that forgiving means excusing. Three different things stacked into one dead end. The rest of this article pulls them apart and offers a way in that runs through neither decision nor conversation: the physical release that comes before forgiveness (in French).

Because it stands guard.

A grudge is a surveillance system. It records who cost you what, it keeps mistrust high, and it stops you handing your arm back to someone who has already bitten it. Seen that way it is neither a moral defect nor a weakness of character: it is a reasonable response to real information. The trouble is not that it exists, it is that it keeps running at full power long after the danger has passed.

An ordinary example. In January 2025, a colleague presents work you prepared, in a meeting, without naming you. Nobody says anything. You let it go, because the moment was ambiguous and claiming it would have looked petty. Twenty months later that colleague has changed companies, and something still tightens when their first name turns up somewhere. The guard stayed at a post that no longer exists.

This is where most attempts fail. People try to talk the vigilance into leaving without giving it a replacement. It refuses, logically. As long as no concrete protection has taken over, the body keeps the grudge because it is still the only security system on duty.

That protection is usually very mundane: stop lending money, space out the visits, stop replying after eight in the evening, tell someone this time. These are not spiritual gestures. They are gestures that make the vigilance redundant, and redundant vigilance eventually falls asleep on its own.

On this point there is a direct measurement, and it is an old one.

In 2001, Witvliet, Ludwig and Vander Laan published a simple experiment in Psychological Science. Seventy-one participants were asked to think about a real offender from their own life, then to hold two mental stances in turn: rehearsing the hurt and nursing the grudge on one side, taking the other person’s perspective and imagining forgiveness on the other. Meanwhile heart rate, blood pressure, skin conductance and the electrical activity of the brow muscle were recorded.

Unforgiving thoughts pushed everything up: more aversive emotion, higher heart rate, higher blood pressure, a distinctly more contracted brow. Forgiving thoughts did the reverse, along with a greater sense of control. Above all, the activation tied to the grudge persisted into the recovery periods, once the exercise was over. The body takes longer to come down than it took to go up.

That recovery detail is the interesting one, because it matches a much broader model. In 2006, Brosschot, Gerin and Thayer set out the perseverative cognition hypothesis in the Journal of Psychosomatic Research: what damages health in psychosocial stress is less the events themselves than the way we keep processing them mentally, before and after. Worry and rumination prolong physiological activation that should switch off within minutes.

An offence lasts three seconds. What lasts years is the rerun.

Put differently, a grudge is not costly because it is unfair or immoral. It is costly because it is a rerun machine, and every rerun calls up a full bodily response to an event that is no longer happening. If your rumination goes well beyond a single person, the more useful angle is the received wisdom about thoughts that keep circling (in French).

One caveat. The 2001 study covers a few minutes of mental imagery in a laboratory, with students, and immediate measurements. It shows a plausible mechanism, not a thirty-year health trajectory.

The literature on forgiveness is larger than people assume, and slightly less spectacular than its reputation.

The reference synthesis is still the meta-analysis by Wade, Hoyt, Kidwell and Worthington, published in 2014 in the Journal of Consulting and Clinical Psychology: 53 post-treatment effect sizes, 2,323 participants, and only interventions explicitly designed to help someone forgive a specific hurt. The result is clear without being startling. People who followed a forgiveness protocol forgave more than people who received nothing, with an average effect around 0.56, and somewhat more than people who followed another approach. Depression and anxiety came down too.

What matters most in that meta-analysis is not the headline effect but its moderators. How long the protocol runs predicts the outcome better than the theoretical model behind it, with modality (individual rather than group) as the other unique predictor. The longer the practice, the better it works, whichever school it comes from. That is usable information: stop hunting for the right method, look for the one you will keep up for several weeks.

The second finding is frankly counter-intuitive. In 2012, Toussaint, Owen and Cheadle followed 1,232 Americans aged 66 and over, representative of the population, and crossed their forms of forgiveness with all-cause mortality. After adjusting for religiousness, socio-demographics and health behaviours, one variable remained associated with higher risk: conditional forgiveness, the kind granted only if an apology or amends arrives first. Part of the effect ran through physical health.

This is an observational study, on an older population, and an association is not causation. But the picture it paints is telling: the costliest position is not open anger, it is waiting. Waiting for an apology means keeping a file open whose closure belongs to somebody else. Plenty of people hold that position for years without noticing they have handed their own relief to the person who hurt them.

Hence a distinction that unblocks a great deal.

Three operations get routinely confused. Excusing means saying the act was acceptable: nobody is asking you to, and it is often untrue. Reconciling means resuming the relationship: that involves the other person, a conversation, sometimes guarantees, and it can be a bad idea. Putting down the load means no longer carrying the tension attached to the memory: it concerns only you, and nobody can stop you.

Only the third is available unilaterally. It is also the only one that lowers what Witvliet and colleagues were measuring. You can go on thinking the person behaved badly, refuse all contact, and still stop clenching your jaw when their name comes up. These two things are not in conflict, they simply live in different drawers.

That leaves the question of the tool. Directed kindness practices, where you form intentions of peace towards yourself and then towards others, were reviewed by Galante and colleagues in 2014: twenty-two randomised trials, moderate effects on depression, compassion and self-compassion against passive controls, inconclusive results against active controls, low to moderate methodological quality. That is an amber light, not a green one. The authors note that this kind of exercise can be difficult at first for some people, which matches experience: forming a kind intention towards someone who harmed you often produces anger before it produces anything calmer.

Which is why the version below asks for kindness towards nobody. If what you are after is addressing yourself after something you blame yourself for, the way in is different and it is described in the compassion phrase ritual (in French).

The exercise takes three minutes and involves no visualisation.

Sit down, close your eyes if that is comfortable, and let the situation come without narrating it. Not the chronology, not the arguments, not what you should have said. Simply look for the place in your body that answers: the jaw tightening, the sternum closing, the shoulders lifting a centimetre, the throat narrowing. There is almost always one, and it is more precise than people expect.

Once you have found it, breathe normally for a few cycles with your attention resting there, without trying to release it. Then, across six to eight out-breaths, place the phrase inwardly: “I let go of what can be let go”. Not “I forgive”, not “I release everything”. The wording carries its own escape clause, and that is what makes it bearable: it commits you only to releasing the part that is ready today, which is sometimes ten per cent.

Close without taking stock. Do not check whether you resent the person less. That question is not due yet.

Two or three minutes a day, five days out of seven, for a few weeks. That rhythm is not decoration: it is what the main finding of the 2014 meta-analysis suggests, where duration weighs more than technique. For the first few days nothing much tends to happen, and plenty of people stop there, concluding it does not work. What changes first is not the feeling, it is how long it takes to settle after a reminder: the scene still comes back, but it takes five minutes instead of an evening.

Some mornings the place you found will not release a millimetre. Those are not failures. They are days when the vigilance judges it still has work to do, and forcing it only confirms it in the role. If you would rather have a step-by-step version, more guided and centred on a single offence, it exists in the six-step protocol for forgiving someone (in French).

This needs saying plainly, because pages about grudges almost always skip it.

The practice is for ordinary grudges: a friendship betrayed, an old family conflict, an injustice at work, a breakup that ended badly. It is not for a situation that is still dangerous. If the person is still doing harm, the priority is not to lower your guard, it is to get out of range: the grudge is doing its job and disarming it would be absurd.

It is also unsuitable when the memory brings on flashbacks, freezing or repeated nightmares. Going back deliberately to a traumatic scene without a frame can worsen symptoms, and support from a professional trained in psychological trauma comes before any breathing exercise. The same caution applies if the rumination sits alongside a low mood that has lasted several months.

One last limit, less serious and more common. A three-minute exercise works on the load of a memory, not on the situation feeding it. If you cross paths with the person daily, if no boundary has been set, if the subject resurfaces at every family meal, the body is right to go back to its post. The practice works far better once something has moved in real life, even if that something is only a polite refusal.

For everything else, the exercise is close to risk-free and needs no equipment. It does not ask you to believe in forgiveness, to like anyone, or to consider what happened acceptable. It asks you to find a place that tightens and breathe on it a few times a day, which is a fairly low bar for a load some people have carried for ten years.