How Do You Save Your Marriage When Only One of You Is Trying

Photo by PNW Production on Pexels

Photo by PNW Production on Pexels
How do you save your marriage is usually asked by one person, on their own, about a situation the other person is not currently treating as an emergency. That asymmetry is the whole problem and almost nothing written on this subject acknowledges it.
Here is the honest structure. One person can start this alone. One person cannot finish it alone. Both halves are true and most advice fails by holding only one of them. The optimistic version says just change yourself and watch them come back, which sets you up to conclude you failed at something that was never fully yours. The bleak version says it takes two, which is accurate about the destination and useless about the next fortnight, because right now there is only you.
So the useful question is not how do I save it. It is what is actually mine to do, how do I do that properly, and how long do I do it before the answer becomes clear on its own.
Worth defining, because people chase two very different things under the same word and only one of them is worth having.
The first is preservation. The marriage continues, the house stays intact, nobody has to tell their parents. That version can be achieved by a person willing to shrink enough, and plenty of people do achieve it. They are still married in ten years and both of them are quietly gone.
The second is repair. The thing that broke it stops, the two of you are actually reachable to each other, and the relationship is one that both people would choose again. That is harder, slower, and considerably less certain. It is also the only one worth spending a year on, and it is worth being honest with yourself early about which one you are pursuing, because the tactics are different and the first one quietly costs more.
Not their behavior. Not their feelings about you. Not whether they go to counseling. What you control is your own conduct, and specifically the parts of it that are currently making the situation worse.
This sounds like a consolation prize and it is not, for a mechanical reason. A marriage in trouble is a loop, and loops require both people to keep running their part. When one person genuinely stops running theirs, the loop cannot continue in the same shape. That is not a guarantee of anything, and it is the only real lever available to a person working alone.
The trap is doing it as a technique. Men in particular will change their behavior for six weeks, notice no response, and conclude the approach does not work. What they were running was a strategy with a deadline attached, and the other person could feel that from the first week. The change has to be one you would keep whether or not it worked, which is the only version that reads as real from the outside.
The most useful research finding in this whole field is that four specific behaviors predict a marriage failing far better than the topics couples argue about. The Gottman Institute calls them the four horsemen, and if you take one practical thing from this page, take these.
These are yours unilaterally. You can stop all four without the other person's cooperation, starting this week, and the effect on a house is noticeable within a month. Nothing else on this page is that cheap.

The instinct when someone is leaving is to argue them out of it. Present the evidence. Remind them of the good years. Point out how much you have changed in the last three weeks. Ask, again, where they are with it.
Every one of those makes it worse, and the reason is worth understanding rather than just obeying. A person considering leaving is trying to hear their own signal. Every argument you make adds noise, and the louder it gets the more they need distance in order to think, which produces exactly the withdrawal you are trying to prevent. You are, with entirely good intentions, pushing them further out.
There is also a limit worth stating plainly. You cannot promise their decision, and neither can anyone else. No program, no book, no amount of work controls what another adult concludes about their own life. Anybody selling you the outcome is selling you something. What you can control is becoming a person genuinely worth choosing, day after day, and then living with the fact that the choosing is not yours to do.
It is the most common description and the least useful one, because it makes the problem sound like weather. Marriages rarely drift. They stop doing two or three specific things, and the drifting is the consequence.
The usual mechanics look like this. Conflict stopped happening, which everybody experienced as an improvement. Both people got very polite. The number of subjects that could be raised shrank quietly, year on year, until the conversation was logistics and the children. Then two people who were technically fine found they had nothing to say to each other on a Saturday.
That is worth naming as a warning sign rather than a safe harbour. A polite marriage with no conflict in it is not peace, it is a void, and it is fertile ground for exactly the things people are trying to avoid. Emotional affairs grow in it, because a person who has not been genuinely reached in three years is unusually reachable by someone new. If that is the shape of yours, the work is not conflict resolution. It is finding out whether either of you can still say a true thing at the kitchen table.
Most attempts at this go badly for structural reasons rather than because either person handled it wrong. They happen at eleven at night, after something set it off, with no agreed end point, covering four years of material at once. Nobody could do well in those conditions.
The version that works is smaller than people expect. Twenty minutes, arranged in advance, one subject, sober, not at bedtime. Your job in it is to understand their position well enough to repeat it back in a way they would agree with. That is the entire job. Not to correct the parts that are unfair, not to supply context, not to explain what you actually meant on the day in question.
This is much harder than it reads, because roughly ninety seconds in you will hear something inaccurate and every instinct will fire at once. Let it be inaccurate. You are not in a dispute about the record, you are finding out how it landed, and how it landed is a fact even when the account of it is not. A person who has been genuinely heard once tends to soften more in that twenty minutes than in six weeks of being reasoned with, and that softening is the thing you have been trying to buy with argument.
Then most of this page is on hold, and it is worth saying that bluntly rather than letting somebody spend a year on the wrong problem.
Nothing works while contact is continuing. Not counseling, not date nights, not six months of impeccable behavior. A marriage cannot be honestly evaluated while one person is running a second relationship alongside it, because what is being weighed is not the marriage, it is a version of it the other person has never actually been shown. If you are the one with the contact, that ends first and completely, and it ends before anything else on this page is attempted.
If you suspect it and cannot confirm it, be careful with what you do next, and be more careful with what you conclude from silence. A marriage where one person is quietly certain and never says it is a marriage running on a fiction that will surface eventually anyway. Say the thing you actually think, once, calmly, without accusing them of specifics you cannot support.
This is the reframe that gets people through the middle of it, and Tyler puts it better than we can improve on. Surrender is not giving up. It is giving over the outcome while still doing the very best you can with what you have.
Those two halves are usually treated as opposites, which is why people oscillate between frantic effort and collapse. Held together they describe the only sustainable posture available: you do everything actually in your power, thoroughly and without a deadline, and you release the part that was never yours. Accept what you cannot change, work on what you can, and get honest about which is which.
Practically, this is what stops the year being unbearable. A man doing the work while auditing his wife's face every evening for a verdict is doing the work in the most expensive possible way. Same actions, half the suffering, if he can put the outcome down.
Not what people hire it for. Couples counseling does not decide this for you and it cannot guarantee the marriage survives, and a counselor who implies otherwise has overstepped.
What it does is get the actual pattern onto the table where both of you can see it, slow the decision down so it is not made inside one terrible week, and, if the marriage does end, make the ending survivable for two people who will share children and a history for the rest of their lives. That last function is undersold and is worth the money on its own.
There is also a format built specifically for the situation where one of you is leaning out and one is leaning in, and it is not couples counseling. Discernment counseling is short, structured, and aimed at producing a clear decision about direction rather than at fixing the marriage. It is frequently the one a reluctant partner will agree to, because it does not require them to commit to saving anything in order to attend. What counseling looks like when a marriage may be ending goes further into how to choose.
Common, and not the end of it. Go on your own. Individual therapy for one partner is a legitimate intervention in a marriage rather than a consolation prize, because you are half the system and changes in half a system propagate.
Ask once, properly, and then stop. Once looks like: I am going to work on this either way, I would rather do some of it with you, and here is a specific appointment at a specific time. Not a campaign. A campaign converts the counseling question into another front in the same argument, and then they are refusing the argument rather than the appointment.
Then let the going be the message. A partner who watches someone attend for four months without being nagged into it is receiving information that no amount of asking would have delivered.
Everything above still applies and two things change.
The first is sequencing. Your own work comes before any joint work, and joint work started too early usually becomes the room where they hear new information, which is the worst possible place for it. If anything is still undisclosed, that gets planned with a professional rather than confessed on a bad night to relieve your own guilt.
The second is that your timeline is not yours. Asking how long this is going to take is a reasonable human question and it lands as pressure, every time. The practical version of all this is set out in how to rebuild trust while living in the same house, which is a different problem from rebuilding it in the abstract.
Six months of doing it properly, and one month tells you nothing. Properly means the four horsemen genuinely gone, your own support in place, the asking stopped, and any ongoing behavior actually stopped rather than paused.
At six months, look at three things rather than at how you feel. Has anything softened in a direction you did not engineer. Have they started making requests of you again, which is the most encouraging signal in this whole territory, because requests mean they are still planning on you being there. And can someone other than you name a change in you.
If all three are flat after six honest months, you have real information. Not a verdict, and considerably more than you had. How to read where the marriage actually is covers what to do with that.
This arrives in most of these situations at some point and it is almost always mishandled, because it lands as rejection and gets answered as though it were an opening bid.
Give it, properly. Properly means actually giving it rather than granting it and then testing it hourly. No checking in to see whether the space is working. No long message on day three explaining that you respect the space. That is not space, it is surveillance with better manners, and they can tell.
The one thing worth negotiating is a shape, because open-ended distance tends to become permanent by default rather than by decision. Something like: we talk on Sundays about the practical things, and we check in properly in a month. That gives them the room they asked for and stops the arrangement quietly becoming the new marriage. Ask for that once, accept whatever version you get, and do not renegotiate it in week two.
Then use the time rather than serving it. Men frequently treat this stretch as a waiting room, come out of it unchanged, and are genuinely puzzled that nothing is different. The space was the opportunity, and what you did with it is the only thing that will have communicated anything at all.
We will not tell anyone whether to stay, and we are not going to pretend every marriage should be preserved either.
If there is violence, threat, or a pattern of control, the question is not how to save it, it is whether you are safe, and that is a different conversation with different help. Short of that, the honest test is whether staying requires you to keep setting down your own values rather than your preferences. Giving things up is what marriage is. Becoming a person you would not recognise, one concession at a time, is not devotion, and no arrangement that requires it is going to hold anyway.
The other version of the wrong goal is subtler and more common. Some people are not trying to save a marriage, they are trying to avoid being the one who ended it, so that the record shows they tried everything. That produces years of effort with no real hope inside it, which the other person can feel and cannot name, and it is a harder way to live than either staying properly or leaving honestly. If that is closer to your actual position, it is worth admitting to yourself first, because the work that follows is completely different.
You will not save this by convincing anyone. The only thing that has ever worked is becoming someone the other person can actually reach, and then staying that person long enough that it stops looking like a campaign. Whether that is enough is genuinely not up to you, and doing it anyway is still the best available use of the year.