What's Your Divorce Language? The Real Signs a Marriage Is Ending

Key Points

Love languages don't have the evidence people assume they do. A 2024 review in Current Directions in Psychological Science found weak support for all three of the framework's core claims. Anything built on top of it inherits that problem.

What predicts a marriage ending is well documented, and contempt leads the list. Gottman's research identified four patterns that forecast dissolution, and contempt is the single strongest of them.

The most predictive signal is also the most boring. Six years after being observed as newlyweds, couples who were still married had responded to each other's small bids for connection about 86 percent of the time. Couples who had divorced: about 33 percent.

Withdrawal is usually late, not early. By the time someone goes quiet, they've often already asked several times and stopped expecting an answer. The silence is the end of a process, not the start of one.

Shutting down often isn't indifference. It's frequently a nervous system past its limit. That distinction changes what you should do about it, and most people get it backwards.

People ask me some version of this question fairly often. What's the language? What's the code? If I could just read the signals correctly, I'd know what's happening in my marriage.

I understand the appeal. A code implies a solution. And "divorce language" is a nice phrase because it borrows the shape of something familiar.

But I'd be doing you a disservice if I built an article on a frame I don't think holds. So let me give you the honest version, which is more useful anyway: there isn't a divorce language, the love-language idea it's modeled on is weaker than most people realize, and there's a large body of research on what actually predicts a marriage ending that almost nobody quotes.

First, About Love Languages

Gary Chapman's five love languages have been a fixture since 1992. Words of affirmation, quality time, gifts, acts of service, physical touch. The premise is that each person has a primary one, and that couples do better when they speak each other's.

In 2024, researchers Emily Impett, Haeyoung Gideon Park, and Amy Muise published a review in Current Directions in Psychological Science testing the three central assumptions against the relationship science literature. None of them held up well.

People don't have one primary language. They generally appreciate all five. The categories overlap enough to question whether they're distinct in the first place. And researchers haven't found a consistent link between relationship satisfaction and whether partners match on languages.

The authors propose a better metaphor. Love isn't a language you have to learn to speak. It's closer to a balanced diet, where a range of expressions meets a range of needs, and no single nutrient carries the whole thing.

I want to be fair to Chapman here, because the idea caught on for a reason. It gives people permission to name a need out loud, and it gives their partner something concrete to do. That's genuinely valuable. The framework is shaky. The impulse behind it is sound.

Which is the same thing I'd say about "divorce language." The instinct to read the room is correct. The map is wrong.

What Actually Predicts It

John Gottman's team spent decades observing couples and following them for years. What they found were four communication patterns that forecast dissolution with unsettling accuracy. They called them the Four Horsemen: criticism, contempt, defensiveness, and stonewalling.

Criticism is different from a complaint, and the distinction matters more than it sounds. A complaint is about an event. "I was worried when you didn't call." Criticism is about a person. "You never think about anyone but yourself." One is solvable. The other requires your partner to defend their character before they can address anything.

Contempt is the one to watch. Sarcasm, mockery, eye-rolling, name-calling, hostile humor, anything delivered from a position of superiority. Gottman identified it as the single greatest predictor of divorce, and it's the pattern people most often fail to recognize in themselves because it hides inside jokes.

Defensiveness is usually a response to criticism, and it functions as a counter-accusation. It reads as "the problem isn't me."

Stonewalling is shutting down, going blank, leaving the room. And this one I want to slow down on.

Shutting Down Is Often Not What It Looks Like

Here's where my own work comes in, and where I disagree with how most articles handle this.

Stonewalling gets read as coldness. As "he's checked out" or "she doesn't care anymore." Sometimes that's true. Often it isn't.

Gottman's team documented a physiological state they called flooding: heart rate up, stress response engaged, and the capacity for empathy and problem-solving offline. Someone in that state isn't withholding. They're overwhelmed, and shutting down is what an overwhelmed nervous system does. Men flood more easily on average, which is part of why the pattern gets gendered.

Why this matters practically: if you read your partner's shutdown as indifference, you'll escalate to get a reaction. That escalation floods them further. You've now built a machine that manufactures the exact thing you're afraid of.

The better read is to notice the state and pause. Twenty minutes, genuinely apart, not spent rehearsing your argument. Come back when you said you would.

I'm not telling you every shutdown is flooding. Some people really have left the marriage emotionally and are running out the clock. I'm telling you that you can't tell the difference by escalating, and the escalation costs you either way.

The Quietest Signal Is the Most Predictive

If you want something that functions like a language, this is the closest thing research offers, and it's better than anything the love-language frame gives you.

A bid is a small attempt at contact. "Look at this." A sigh. A hand on your shoulder. "How was your day." They're constant and mostly unremarkable, and every one of them carries the same underlying question, which is roughly: are you there?

You can respond three ways. Turn toward, meaning you engage, even minimally. Turn away, meaning you miss it or ignore it. Turn against, meaning you snap.

Gottman observed newlywed couples and followed up six years later. Couples who were still married had turned toward each other's bids about 86 percent of the time. Couples who had divorced: about 33 percent.

That gap is the whole thing. Not fighting. Not values. Not chemistry. The mundane response rate to mundane requests for attention.

So if you want a signal to watch, watch this one. Not whether your partner is affectionate. Whether the small offers still get answered, in both directions. And notice something uncomfortable: the number that predicted divorce was 33 percent, not zero. These marriages didn't have no connection. They had roughly a third of enough.

Reading the Signs, Reordered

The usual list of warning signs mixes strong signals with weak ones. Here's how I'd rank them.

Contempt, in any amount. The strongest predictor there is, and it usually gets normalized as "that's just how we talk to each other."

Bids that stopped. Not bids that get refused. Bids that stopped being made. Someone who has quit asking has usually asked plenty of times already.

A shutdown you can't read. Worth investigating rather than assuming, per the section above.

A separate financial life forming. New accounts, large solo purchases, a sudden interest in credit. This one is genuinely predictive and it's also the one with legal weight, so it's worth taking seriously in a practical way rather than only an emotional one.

Life reorganizing around a future you're not in. Not new hobbies, which are fine and healthy. Plans that quietly stopped including you.

One item from the standard list I'd downgrade: less physical affection. It's real, but it moves for a dozen reasons that have nothing to do with divorce. Illness, medication, exhaustion, a new baby, depression, age. On its own it tells you much less than people think.

If You're Seeing These, Do Two Things

Not one. Both, even if you're hoping to stay.

The relational one

Don't open with an accusation, and don't open with the list you just read. "I think you want a divorce" is a conversation-ender.

Open with a single specific observation and a real question. "I noticed we haven't talked about anything but logistics in a while. Is something going on for you?" Then stop talking and let the silence sit longer than is comfortable.

And time it. You cannot have this conversation while either of you is flooded, and you cannot have it at 11 p.m. after a bad day. Pick a moment when both nervous systems are actually available. This is the least romantic advice in the article and it's the piece that most determines how the conversation goes.

If you get a real answer, whatever it is, that's a good outcome. Even a hard one. Information is what you came for.

The practical one

Knowing where you stand is not disloyalty and it isn't a decision. It's how you stop being surprised.

Know your numbers. Locate the account statements, the tax returns, the retirement balances, the deeds and titles. Have a credit card and a bank account in your own name, because credit history takes time to build and you don't want to start from zero later. And consider one consultation with a family law attorney. A single meeting to understand what the process would look like in your state commits you to nothing.

If you're in Louisiana, there's a specific reason not to wait, and it's one people learn too late. Under Civil Code Article 103.1, a no-fault divorce requires living separate and apart for 180 days if you have no minor children and 365 days if you do. The clock doesn't start when you decide. It starts when you're actually living in different residences, and reconciling during that period resets it. So in Louisiana, "let's see how the next few months go" has a price measured in months on the back end. Other states vary widely, which is exactly why one consultation is worth it.

The Reframe

Stop looking for a code. There isn't one, and the search itself tends to keep you in your head instead of in the room.

Watch what happens to the small stuff. Whether the offers still get answered. Whether there's contempt in the joking. Whether the shutdown is a wall or a person over their limit.

Those are readable, they're documented, and unlike a love-language quiz, they point at something you can actually do this week.

Frequently Asked Questions

Is there really such a thing as a divorce language? Not as a research concept. It's a riff on love languages, and a 2024 review in Current Directions in Psychological Science found weak support for that framework's core assumptions. What does exist is a well-studied set of behavioral patterns that predict marriages ending, which is more useful anyway.

What is the biggest predictor that a marriage will end? Contempt. Among Gottman's Four Horsemen, it's identified as the single strongest predictor of divorce. It shows up as sarcasm, mockery, eye-rolling, and hostile humor, which is why it often gets dismissed as a sense of humor rather than recognized as a pattern.

How can I tell if my spouse wants a divorce? Look at bids for connection rather than at moods. Couples still married six years after being observed had turned toward each other's small bids about 86 percent of the time; couples who divorced, about 33 percent. Also watch for contempt, for a financial life forming separately, and for plans that stopped including you.

My partner shuts down during arguments. Does that mean they've given up? Not necessarily. Stonewalling is often flooding, a physiological state where the capacity for empathy and problem-solving shuts off. Escalating to get a response makes it worse. Take twenty minutes apart without rehearsing your case, then return. If it happens constantly regardless of how calmly you raise things, that's worth exploring with a professional.

Should I talk to a lawyer if I only suspect my spouse wants out? A single consultation commits you to nothing and tells you what the process would involve where you live. That's especially worth knowing in states with long separation requirements. Louisiana, for instance, requires 180 or 365 days of living separately under Civil Code Article 103.1, and that clock starts when you physically separate, not when you decide.

What should I do first if I see these signs? One conversation and one set of practical steps. For the conversation, lead with a single specific observation and a real question, at a time when neither of you is worked up. For the practical side, locate your financial documents, make sure you have credit in your own name, and consider one attorney consultation. Doing the second doesn't mean you've given up on the first.