Why does a relationship ending hurt in my body, and how long does it actually last?

Why Are You So Much Worse at Predicting This Than You Think?

This is the most useful and most checkable fact on the page.

People systematically overestimate how badly and for how long a bad event will affect them. That general effect has been established across many domains and is discussed elsewhere on this site in the context of predicting how much you will enjoy an evening you want to cancel. What is more directly relevant here is that somebody went and measured it on this exact event.

Krishnamurti and colleagues published a study in 2008 that did something unusually clean. They took undergraduates who were currently in relationships and asked them to forecast how distressed they would be at various points if the relationship ended. Then they waited. The ones whose relationships actually ended were followed with repeated distress measures over the following months, so the forecasts could be compared against the same people's later reported experience.

The forecasts overshot. Real distress was genuinely present and it declined faster than the same people had predicted it would, with the gap between prediction and reality widest at the later time points. They were roughly right about how bad the first stretch would be and substantially wrong about how long it would stay that bad.

I should say plainly what I have done with the attribution, because it is unusual. The study has four authors, and three of them are cited on other pages of this site for entirely unrelated work. Naming a co-author rather than the first author is a deliberate choice to stop a set of pages from looking like it rests on four names, and anybody who wants the paper will find it from the description.

And the honest limits. The sample was undergraduates at one university, and their relationships were short by adult standards. Nothing about that design tells you what happens after a nineteen-year marriage. The direction of the finding is robust and has held up across a lot of related work. The popular version, which turns it into a promise that you will be fine sooner than you think, is doing something the study cannot support. A prediction being too pessimistic on average says nothing about where any individual sits.

What it is legitimately good for is this. The sentence that runs at quarter past four in the morning, the one that says this is permanent, is a prediction, and it is being generated by the machinery known to be worst at exactly this. That does not make the feeling wrong. It makes the forecast unreliable, and the forecast is the part doing most of the damage.

What Is the Popular Advice Here That Is Genuinely Dangerous?

One thing on this topic is worth naming as unsafe rather than merely unhelpful.

Because of the imaging work described above, a piece of advice has spread widely which suggests that since emotional pain runs on machinery shared with physical pain, you might reasonably reach for the ordinary over-the-counter painkiller in your kitchen cupboard when an ending is at its worst. It appears in articles, in videos, and in the sort of list that circulates with no source attached. There is a real experiment behind it, in which participants took a daily dose over a few weeks and reported slightly lower hurt feelings on a questionnaire than a comparison group did.

I am not going to name the compound, and that is deliberate, because naming it is exactly how this advice travels.

Here is why it should not be followed. The reported effect was small and measured on self-report scales, which is a very long way from establishing that somebody in the middle of an ending should be dosing themselves. The compound most commonly named in this advice has an unusually narrow margin between an ordinary dose and a harmful one, and it is already among the most common causes of serious accidental harm in its class. The population most likely to act on the advice is a population sleeping badly, eating erratically, sometimes drinking more than usual, and not tracking what it has taken. There is also work suggesting these compounds blunt positive responses along with negative ones, which means the best available case is a flattening rather than a relief.

If any part of this is bad enough that medicating it seems reasonable, that is a conversation with a doctor rather than a decision made at quarter past four in the morning on the strength of something you read.

Why Do I Feel Like I No Longer Know Who I Am Without My Ex?

Some of what you would have said about yourself was only ever true inside the pairing, and it left when they did. The same mechanism is why practical life feels unmoored out of all proportion to how much you miss them.

Because the person was also a structure, and the structure went at the same moment.

This is the most under-described part of an ending, and it accounts for a large share of the disorientation people cannot explain to themselves. You can be entirely clear that the relationship needed to end, feel relatively little about the person on a given Tuesday, and still find that you cannot decide what to do with a Saturday, cannot work out the finances, cannot remember how you used to make plans, and stand in a store holding two things with no way of choosing between them.

What has actually gone is a daily architecture. A schedule that was set by somebody else's schedule. A division of small competencies, some of which you never learned because you never had to. A route into other people, because a substantial part of most adults' social life runs through a partner. A default answer to the question of what happens this weekend, at the holidays, in five years. And a picture of the future, which is the item people underestimate most, because it was doing constant quiet work and was never once examined.

There is also a self-description in there. Slotter and colleagues reported in 2010 on what happens to how people describe themselves after a breakup, finding that those reporting the largest change in their own self-concept also reported the most emotional distress, along with a marked confusion about who they now were. That work is correlational, it rests on self-report, and its samples were students, so it establishes an association rather than a cause. It matches what people describe closely enough to be worth knowing.

The practical consequence is worth stating. If most of what you are experiencing is architectural, then the things that help are architectural too, and they have very little to do with feelings about the person. Somebody has to decide when you eat now, and it has to be you. Somebody has to fill a Saturday. This is unglamorous, it feels beside the point, and it moves things more than most of what gets recommended.

Why Do I Keep Checking My Ex's Profile Even Though It Makes Me Feel Worse?

Because checking sits in the chosen half of an aftermath whose other half is involuntary, and the chosen half keeps running because it pays out faster than it costs.

This is the most useful cut available and almost nobody makes it, which is why the shame in this area is so badly distributed.

The involuntary part is nearly everything described so far. The physical drop at a notification. Waking at the same hour. Appetite. The state that arrives whole when a song starts. The pull toward the person. None of this is being chosen, none of it responds to being argued with, and it will run largely on its own course whatever you decide about it. Feeling ashamed of this part is like being ashamed of a bruise.

The chosen part is much smaller and much more specific. It is what you do in the twenty minutes after waking at four. Whether you open the exchange again. Whether you check. Whether you ask a mutual friend something you have carefully made sound casual. Whether you go past a particular building on a route that is eleven minutes longer. Whether you construct, for the ninth time, the conversation in which they finally understand.

Almost every piece of advice you will be given aims at the first category, which is why almost all of it fails. You cannot decide to stop missing somebody. You can decide, with difficulty and imperfectly, what to do at quarter past four, and that is the entire available surface.

It is worth being accurate about why the chosen part persists, because willpower is the wrong frame for it. Each of those acts pays out. A check produces about twenty seconds of relief from not knowing. A reconstructed conversation produces a version of the ending in which you are understood. A drive past a building produces the sensation of contact. These are small, reliable, immediate returns, and the cost lands ten minutes later, which is the exact structure of every behavior that is hard to stop. Nothing about you is weak. The schedule of payoffs is simply arranged badly.

There is one study pointed straight at whether it makes things worse. A 2012 report on people who kept watching a former partner's profile after an ending found more distress, more longing and less of the personal growth that usually follows an ending, after accounting for how much offline contact those people had. It is named and discussed on the page about checking somebody's social media, so I have described it here without repeating the attribution, and its limit is the one that matters most to you: it measures an association across people at a single moment, so it cannot say whether the checking produced the distress, the distress produced the checking, or both at once.

Which is the honest reason this page hands you no rule about contact, no day count, and no set of phases. Those exist in quantity, they sell well, and there is nothing underneath them. What is worth having instead is an accurate account of what your own four in the morning is buying.

What If This Is an Ordinary Ending Running Its Ordinary Course?

Then nothing is wrong, and being told so precisely is the most useful thing available.

This is the majority position and it is chronically under-described, because a description of the ordinary does not attract attention the way an explanation does. So here it is at length.

Something real ended. It had been running long enough to be built into your week, your body and your plans. The ending removed it at a stroke, and every habit built around it is now firing into empty space. Your appetite went. Your sleep broke. Your attention narrowed onto one subject with a persistence you have never experienced about anything else. Your practical competence dropped. You did some things at four in the morning. Your body reacted to a message from the dry cleaner as though it were a threat.

Every item on that list is what happens. Not what happens to people who are fragile, or who were too invested, or who have unresolved things. What happens.

And it has a course. The loud stretch. The dip when the scaffolding comes down. The long intermittent tail where cues still work but the states get shorter. The morning where it turns out yesterday went by without it. You are not steering any of that, and the parts of you currently trying to steer it are producing more suffering than the ending is.

Two markers fit this root. The first is that the distress attaches to the ending rather than to you. When you look at it honestly, the sentence underneath is about them being gone rather than about what their leaving proved about your value. The second is that it moves. Not steadily and not fast, but a month is measurably different from the previous one in some direction.

What people in this root usually need is permission and a schedule. Permission to stop assessing their own reaction, and a set of concrete external structures for eating, moving and being around other people, none of which treat the loss and all of which make the hours survivable while the involuntary part does what it does.

And the thing worth saying to somebody here, which almost nobody says, is that the size of this is not a problem to be solved. A large response to a real loss is a working system doing its job. The problem, where there is one, is usually the second layer, which is the running commentary about whether the response is appropriate. Take that layer away and what is left is grief, which is unpleasant and finite and does not need your management.

What If This Ending Is Carrying an Older One?

Then the size will keep outrunning the event, and every attempt to make it proportionate will fail, because proportion was never what it was about.

The signature is a mismatch you can measure. A relationship of eight months producing a response that a much longer one did not. A response that arrived at a specific moment, often the moment somebody physically left a building, and that felt much older than the situation. A quality in the distress closer to abandonment than to sadness, arriving with a certainty about being left that was fully formed rather than reasoned into.

Another marker is that the content drifts. People in this root notice that the four in the morning material is often about a parent, an earlier ending, a childhood house, or a period they had considered settled, rather than about the ex-partner at all. The current ending opened a door and something else came through it.

This is genuinely different work and it is not done by understanding the current relationship better. What tends to help first is separating the two, so that the current ending is allowed to be its own size and the older thing is allowed to be its own subject. The page on grief that never got grieved goes into the older half properly, and a reader who recognized themselves in the paragraphs above will get more from it than from the rest of this page.

Why Do Time Heals, No Contact and Getting Back Out There Not Appear on This Page?

Because each of them is either untestable, oversold, or a verdict wearing the clothes of advice.

Time heals describes an outcome and hands you nothing to do with the next hour, which is the only unit that exists to somebody in week two. It is also only half accurate. The involuntary part does fade largely on its own. The chosen part, the checking and the reconstructing, keeps the whole thing refreshed for as long as it runs, which is why two people can wait the same number of months and arrive somewhere completely different. Waiting is not the active ingredient, and telling somebody that it is leaves them with nothing but endurance.

The contact rules are the biggest industry in this territory and there is very little under the specifics. Day counts, phase structures and staged protocols get presented with a confidence no evidence supports. The one mechanical thing that is true is unremarkable: fewer reopenings per day means fewer reopenings per day. Whether reducing contact is even available to you depends on children, housing, work and shared obligations, none of which appear in the protocols. So this page describes what the checking is buying and leaves the arrangements to you, because you are the only person who knows them.

Getting back out there and its opposite, the warning about rebounds, are both verdicts issued in advance. People form new attachments at every possible distance from an ending and the outcomes vary enormously. The useful question is whether what you want is company or evidence, rather than how long it has been, and that one you can answer honestly in about four seconds.

Self-care as usually packaged belongs in the same list. What actually helps in the first month is unglamorous and physical, it does not treat the loss, and it is closer to keeping a house warm than to healing anything.

Where Does This Stop and Something Else Begin?

The markers are practical, and the useful ones are behavioral rather than emotional.

Whether you are eating and sleeping at all, as against badly. Whether the flatness has stayed attached to this subject or spread into everything. Whether there has been any movement in any direction over months. Whether the thinking has become entirely about your own worthlessness rather than about the loss. Whether any thought of harming yourself has appeared at all.

Any of those means the boundary section near the top of this page applies, and the sentence in it is the whole of my advice on it. A page can describe a shape. It cannot assess you, and it should not try.

The other version of this question is duration with none of the markers above. Somebody functional, eating, working, and still substantially organized around an ending three years later sits outside the territory this section is about. Duration like that usually means one of the roots has gone unaddressed rather than anything escalating, most often the older loss or the identity one.

What Does an Ordinary Week Look Like on the Other Side?

Less dramatic than people imagine, which is the part worth knowing in advance.

The feeling does not disappear. People expect a state of not caring and that is not what arrives. What arrives is that the person becomes somebody you knew well and no longer live around. You can think about them deliberately and feel something, and the something has an ordinary size and a short duration.

The cues keep working. A song still does it. What changes is duration. What used to take an afternoon takes twenty-three minutes, then takes as long as it takes to cross a road, and eventually is over before you have decided how to feel about it.

The mornings come back first for most people, and it is a surprisingly precise marker. There is a stretch where the first conscious thought each day is the ending, and then a morning where the first thought is something ordinary and irrelevant, and you notice the substitution afterward with something close to suspicion.

The practical architecture gets rebuilt and is mostly invisible. Saturdays have something in them. You have learned the competencies that were theirs, badly at first. Trash night is solved. Nobody congratulates anybody for this and it accounts for a large share of how much better the weeks feel.

And the four in the morning behavior stops, usually without a decision being made about it. What people describe is not a successful act of restraint but a gradual discovery that the check has stopped paying out. There is nothing new to find, the twenty seconds of relief no longer arrives, and the habit dies of malnutrition rather than of willpower.

How Long Does This Take, Honestly?

It depends on the root, and I will not give you a number for the first one, for the reasons set out above.

The one thing anybody has measured directly against this question is a forecast rather than a duration. The 2008 study in the Journal of Experimental Social Psychology described earlier compared what people predicted about their own distress with what those same people later reported, and the predictions ran both too long and too dark, with the gap widest at the later points. Its sample was undergraduates in short relationships, so it fixes a direction and leaves the duration open, and it cannot tell you where your own weeks will sit inside it.

If this is an ordinary ending on its ordinary course, the honest answer is that the direction is reliable and the duration is not. What is available immediately, on the day you read this, is the removal of the second layer. A great many people find that the commentary about whether their reaction is acceptable was carrying half the weight, and that half can go this week.

If a structure ended, this moves at the speed of the rebuilding rather than the speed of the feeling, which makes it unusually responsive to deliberate action. People who fill the architecture on purpose report a different month within about six weeks. People who wait for the feeling to change first tend to wait a long time.

If an older loss has been reopened, the current ending will often settle on its own schedule while the older thing stays exactly where it was. That part does not resolve through this ending and it is not measured in weeks.

If the difficulty is identity, that is slow and it moves through action rather than insight. Months rather than weeks, and the movement is usually invisible until somebody asks you a question about yourself and you notice that you had an answer.

If what you have is relief, there is frequently nothing to wait for at all, and the only thing that needed to happen was somebody telling you that this is common.

What Is Worth Doing This Week, Before You Decide Anything?

Two things, and neither of them is aimed at the feeling.

The first is a schedule rather than a mood. Three items only. A time you eat, whether or not you want to. A daily physical thing that gets you outside, however short. One arrangement a week that puts you in a room with other people while you are poor company. Write the three down with actual times attached. None of this treats the loss and none of it is supposed to. It changes how survivable the hours are while the involuntary part runs, and it is the only part of the first month that reliably repays effort.

The second is a count, and it is the one that produces information. For seven days, keep a note on your phone. Every time you check, reread, drive past, ask a mutual friend something, or construct the conversation in your head, add a line with the time and one word for what happened just before it. Change nothing else. Do not try to reduce the number. The count is the point.

At the end of the week two things will be visible that were not visible before. The first is the total, which is almost always larger than people expect and is the most persuasive argument anybody will ever make to you about this. The second is the pattern in the word before it, which is where the actual finding lives. Most people discover that their four in the morning is triggered by something specific and repeatable, and that the trigger has very little to do with the ex-partner. Being alone at a particular hour. A particular kind of day at work. The gap after finishing something. One person's name coming up.

That is a piece of information about your own week that no article can hand you, and it is what makes the chosen part of this addressable at all.

One finding bears on the count itself, and it is an odd one. A 2015 study in Social Psychological and Personality Science, whose title reports that participating in research on romantic breakups promotes emotional recovery via changes in self-concept clarity, put 210 recently separated young adults into either four rounds of detailed measurement across nine weeks or an assessment at the start and the end only, and the intensively measured group reported a larger fall in self-concept disturbance, with that gain in self-description accounting for their lower emotional intrusion and loneliness. No other main effect appeared, the mechanism is inferred rather than shown, and nobody in it was told the measuring was meant to help, so treat it as a reason the count is worth keeping rather than as a promise about what it will do for you. Its authors are cited on other pages of this site, one of them for entirely unrelated work, so I have described the paper without adding their names here.

Do not make any large decision this week. Not about a job, a city, money, or another person. The forecasting problem is at its most active in exactly this stretch, and the sentence that runs at four in the morning about how permanent this is comes from the machinery known to be least reliable about precisely that question.

Our Sacred Valley seminar works at the layer underneath the count. We do not start with what you did at quarter past four, because you can already describe it and will have just written seven days of it down. We start with what it has been buying, since a behavior that repeats that reliably at that cost is never doing nothing, and until you can see the payout, every resolution to stop will fail in the same place at the same hour. Then we look at which of the five roots is actually yours, because a structural problem treated as grief and an older loss treated as a recent one both waste years, and we build a plan you are held accountable to afterward, back in your own week with the same quiet Sunday afternoons in it. The goal is not to feed your emotions. Everybody is screened before arriving, which keeps the group small enough that somebody in the middle of an ending gets worked with properly rather than handed a worksheet. One limitation worth stating plainly: exact dates, the price, and the final cap on group size are still being settled, so anybody comparing programs on those numbers this week cannot yet compare ours. Ask every program you are weighing the same questions, this one included, and decide from what comes back.

The count costs nothing, takes seven days, and is worth having whichever way it comes out.