How do you keep a couples agreement on the night you are too upset to want to keep it?

What each one is actually for

Secure Functioning

Most couples run on an implicit deal nobody has ever stated, and discover its terms only when one of them is breached. Secure Functioning is Tatkin's proposal that you state the deal instead. In his PACT model — a Psychobiological Approach to Couple Therapy, so named because it treats the pair as two bodies that regulate each other rather than as two arguers exchanging positions — partners write explicit agreements about what they owe each other and what is out of bounds: we do not threaten to leave in a fight, we do not undermine each other in front of other people, we do not let a rupture stand overnight without contact. Around those he builds small compulsory rituals at the moments the pair is most exposed — the parting in the morning, the first two minutes of the reunion, the last exchange before sleep. He calls the protected unit the couple bubble. Writing the terms down while nobody is upset is not a matter of tidiness: the capacity to be fair and generous is precisely the capacity that goes offline in a fight, so nothing you would need it for can be invented at the time.

Acceptance and Commitment Therapy

ACT starts from an unusual premise: the problem is rarely the thought or the feeling itself, it is what you are doing to get away from it. Steven Hayes and colleagues built it around changing your relationship to internal experience rather than the experience's content. Three moves matter for what follows here. Defusion means seeing a thought as a thought — a piece of language passing through, not a fact and not an order — which in practice can be as small as adding a prefix: not they have stopped caring but I am having the thought that they have stopped caring. Acceptance means allowing a feeling to be present in the body without fighting it, suppressing it or acting on its urge, on the finding that the struggle against an unpleasant state reliably intensifies the state. And committed action means picking the values you actually want to live by and behaving accordingly now, in the presence of whatever is going on internally, rather than waiting for the internal weather to improve first. That last one has the counter-intuitive edge: in ACT, motivation is treated as something that arrives after the action, not as its prerequisite.

What you get out of pairing them

What Secure Functioning gives you is a relationship that does not depend on either person being at their best, which is an enormous relief the first time you feel it work. What ACT gives you is the ability to act on something you chose while feeling something you did not choose. Combine them and the payoff is narrow and specific: the agreements start surviving the nights they were written for. That is worth naming precisely, because it is not a general improvement. Most couples who make agreements find they hold beautifully in ordinary weeks and evaporate on the two or three occasions a year when they would actually have made a difference — the night someone comes home already raw, the argument that starts at 11pm, the week after a loss when nobody has anything to give. The failure is not bad faith and it is not a weak commitment. It is that in those states people cannot locate the agreement, or can locate it and cannot want it, and a rule that requires wanting is a rule with a hole in it exactly where the load is. ACT's contribution is that wanting was never required. There is a second, quieter gain in the other direction: private psychological work is unverifiable even to the person doing it, and a partner who watches you name a thought instead of firing it gives you the only external evidence available that anything changed.

What Secure Functioning rests on

• Attunement requires active monitoring not good intentions clinical practice
  Staying attuned to a partner's internal state requires ongoing structured attention rather than goodwill alone because connection erodes by default without it.
• Explicit commitment changes daily priority weighting clinical practice
  Explicitly and mutually committing to protect the relationship above other priorities changes how partners weigh choices during ordinary daily conflict.
• Rituals of connection reduce reliance on spontaneous attunement practitioner report
  Building repeatable rituals of connection reduces dependence on spontaneously noticing a partner's needs in the moment.

What Acceptance and Commitment Therapy rests on

• Acceptance reduces struggle randomised trials
  Allowing an internal experience to be present without resisting or acting on it reduces the psychological struggle that resistance would otherwise intensify.
• Decentering from thought randomised trials
  Shifting attention from the content of a thought to observing it as a passing event in awareness creates distance from that content instead of fusion with it.
• Behavioral activation precedes motivation randomised trials
  Taking action before motivation appears, rather than waiting for motivation to arise first, itself generates the motivation and improved mood that waiting would delay.
• Values clarification guides action clinical practice
  Identifying personally chosen values gives behavior a stable compass that does not depend on momentary emotion.

Where they actually meet

These two models do not share a principle in our records, and the reason is instructive: they are working on different objects that happen to be stacked on top of each other. Tatkin's object is the pair. His model is trying to take something that was invisible — a partner's internal state, monitored by nobody, assumed rather than checked — and make it a thing both people actively track, held in shared attention. ACT's object is one person's inside. It is trying to take something you were living inside, fused with, acting from — you are your anxious thought, it is not a thought, it is simply what is true — and make it something you can watch pass through. Same manoeuvre, two different materials, and put side by side the seam becomes obvious: a couple can be doing excellent mutual monitoring while one partner is still completely fused with the thought that is generating the state being monitored. The attunement has nothing to attach to. Your partner can see that you are upset, can say so warmly, can offer exactly what your agreements promise, and none of it lands, because from inside the fusion their warmth is being processed as evidence in a case you have already decided. What ACT loosens is the fusion, which is the precondition for anything relational to get through. And what PACT provides, going the other way, is the thing the defusion work otherwise lacks entirely: a witness. Whether you truly held a thought lightly or merely told yourself you did is not a question you can answer from the inside, and the person across the table is the only instrument available.

The principle underneath both

An agreement that only holds while you feel like keeping it was never doing any work. That is the pattern under both practices, and it belongs to neither — it is the logic of a vow, of a contract, of a training programme you keep on the wet Tuesday, of every arrangement human beings have built precisely because they expect their future selves to be unreliable. Both models are unusual in taking that expectation seriously rather than treating it as pessimism. Tatkin builds the expectation into the structure: the agreements are written when calm because he assumes calm will not be available later, and the rituals are compulsory because he assumes nobody will spontaneously feel like doing them. ACT builds it into the individual: you do not need the feeling to change first, you need to be able to carry the feeling while you act, which is a completely different skill from feeling better and is trained differently. Put them together and you have the same insistence applied at two scales — a two-person arrangement designed for bad nights, and a one-person skill for the bad night itself.

What that makes the real goal

Behave like the person you said you were while the feeling that argues otherwise is still in the room. That is the aim, and stating it that way frees it from both traditions at once. It is not emotional regulation, because nothing about the feeling has to change. It is not communication skill, because you may say very little. It is the capacity to let a chosen commitment govern behaviour under internal conditions that are actively voting against it — which is what makes it a training target rather than a personality trait, since the way to get better at it is to practise it at low stakes until the machinery is familiar. Once the goal is put like that, other traditions become visibly relevant that would otherwise look like they belong to a different shelf entirely, and the relationship case turns out to be a hard instance of a general problem rather than a special subject.

Other practices that serve the same goal

Emotionally Focused Therapy, developed by Sue Johnson, comes at the same objective from the opposite end and has stronger couples-specific trial evidence than anything else on this page. Where ACT works on your relationship to your own internal experience, EFT works by getting the vulnerable feeling under the reactive one to be spoken aloud to the partner and responded to — the objective is the same, acting from what you actually care about rather than from the protective reflex, but the mechanism is disclosure to another person rather than private defusion. Second, Nonviolent Communication, Marshall Rosenberg's four-part structure of observation, feeling, need and request, which gets at the same target by constraining the sentence you are allowed to say until the reactive content has been separated out of it. Its evidence base is much thinner — practitioner reports and small studies rather than substantial trials — but as a piece of scaffolding for a heated moment it is concrete in a way ACT's internal moves are not, and some people find a required sentence structure far easier to use under pressure than an internal manoeuvre. Third, mindfulness-based programmes adapted for couples, which train the attentional groundwork both other routes assume you already have; the trials there are small and the outcomes are usually satisfaction measures rather than behaviour under conflict, so treat them as promising rather than established.

A tradition that already does this

This path has already been walked, and by people who were explicit about it. Integrative Behavioral Couple Therapy, developed by Andrew Christensen and Neil Jacobson, took an existing behavioural couples model built around change — negotiate, agree, do differently — and added acceptance work as a co-equal component, on the observation that some differences between two people are not going to be negotiated away and that the struggle against them was doing more damage than the differences themselves. That is the same argument ACT makes about a thought or a feeling, arrived at independently and applied to a partner instead. It was tested against traditional behavioural couple therapy in a large randomized trial, which makes it the only thing in this vicinity with that grade of evidence behind the acceptance half. Its existence is the strongest reason to think the pairing described on this page is a sound shape rather than a clever juxtaposition — though note carefully what it is not: IBCT is not PACT, and it is not ACT, and its trial says nothing directly about either.

Using Secure Functioning (PACT) to enhance ACT

ACT's defusion and acceptance work is usually done alone or in a therapist's office, which means the only witness to whether it actually happened is the person with the strongest interest in believing it did. Tatkin's model supplies a second nervous system in the room and a structure for using it, and there are two concrete forms worth running. The first is a witnessed naming, built into a fixed check-in rather than left to spontaneous conversation — same time, twice a week, ten minutes, non-negotiable. Each of you names one thought that had you that day, said aloud in ACT's own grammar: I had the thought that you were pulling away, and I am not treating it as a fact. Your partner's job here is precise and counter-intuitive, and it is where couples get it wrong. It is not to reassure you, and it is not to argue with the content. Under PACT's attunement principle their job is to register, visibly, that the thought got named rather than acted on — a nod, an okay, I heard that, maybe a hand. Reassurance would treat the thought as a claim requiring a verdict, which fuses it right back on. Visible registration treats it as an event that happened in you and was survived, which is the thing solitary practice cannot give itself. The second form is for the values half. ACT asks you to clarify what you actually care about; PACT asks two people to protect something jointly. Put them together on a Sunday: each of you writes three values in your own words — not admirable nouns, plain phrases like being someone she can tell bad news to. Then translate each into one action this week that the other person would actually notice, and hand the list over. The handing over is the mechanism. A private value is unfalsifiable and drifts towards whatever you did anyway; a value with an observable action, held by someone who will see whether it happened, stops being a self-description and becomes a commitment. Ask your partner to tell you when they saw it, not whether you were good.

Using ACT to enhance Secure Functioning (PACT)

PACT's agreements are written by people who are calm, and the moments they matter most are moments when at least one partner is not — heart rate up, thinking narrowed, the old protective reflex running the show. That is where an explicit commitment quietly stops working. The person did not break the agreement in bad faith; they lost access to it. ACT contributes three things to that gap, and the third is the one people miss. First, acceptance rather than analysis: instead of thinking your way out of the flooded state or judging yourself for being in it, you let the surge of fear or anger sit in the body without acting on its urge to attack or withdraw, for the length of a couple of slow breaths, before you speak. Build the line into the ritual itself — I'm flooded, give me a minute — said the instant the urge arrives, and understood by both of you as permission to let the state pass through rather than as an exit from the conversation. Second, defusion, in the form of a spoken prefix. Saying I'm having the thought that you don't want to be here out loud is a different act from saying you don't want to be here, and the difference is audible to your partner as well as to you; it converts an accusation into a report, which is the only form in which that content can be said without starting a second fight. Third, and this is the piece Tatkin's model has no mechanism for: do the ritual before you want to. ACT holds, with trial support from behavioural activation research in depression, that action comes first and motivation follows, rather than the reverse. Applied here it is blunt and useful — on the night you least feel like the two-minute check-in, that is when it goes ahead, at its usual time, in its shortened form, with nobody pretending to feel warm. You are not faking. You are doing the agreed thing while feeling what you feel, and treating the wish to do it as an output rather than an entry requirement. Most couples' rituals die on exactly the nights this rule covers, which is why adding it does more than adding another agreement would.

Ten minutes before they walk in, and you are already rehearsing the fight

You are standing in the kitchen and the thought arrives fully formed: they are going to be distant tonight, they always are on Thursdays. Your chest tightens and the old plan starts up — get ahead of it, say the sharp thing first, at least then you are not the one waiting to be disappointed. This is the moment both models are actually for, and they go in a particular order. ACT first, because nothing relational will land until this part is done: you notice that the sentence is a thought and not something you have confirmed, you say it to yourself in the longer grammar — I am having the thought that they will be distant — and you let the tightness sit there for one breath instead of discharging it into the pre-emptive line. The tightness does not go away. That is not a failure of the technique; that is the technique, and expecting it to dissolve is how people conclude this does not work for them. Then PACT, when they come through the door: you use the two-minute check-in you built together, at the time you built it for, and you say it out loud — I had the thought you would be distant tonight, and I nearly came at you about it before you sat down. You are not asking them to disprove the thought. You are letting them see that you caught it. Their job, which you agreed on months ago, is not to defend Thursday or to reassure you; it is to register that this happened, and what they say is some version of thank you for telling me instead of firing it at me. Now the part that makes this a Thursday habit rather than a good evening. You did not feel like doing the check-in. You were still braced when you started it. You did it anyway, at the usual time, in its short form, and about ninety seconds in the bracing dropped by itself — which is the sequence ACT predicts and the opposite of the one you would have waited for. That sentence only worked because the thought got loosened first; without the defusion it comes out as an accusation, and without the ritual it stays a private victory nobody else ever benefits from.

When this is worth reaching for

The agreement holds on ordinary weeks. It went missing on the one night it was built for — the night that started at eleven, or the week after the funeral, or the evening one of you came home already raw and the other could not reach them. Those are the conditions. A second: the check-in is technically still happening and has gone hollow, both of you performing a ritual neither of you is inside, which usually means the ritual is running on a night when at least one of you is fused with something they have not said out loud. A third, and quite different: one partner is already in ACT therapy or has been reading about it, and the vocabulary has started arriving in the relationship without an agreement about how it will be used — which is worth getting ahead of, because those terms can be used to open a conversation or to hold one at arm's length and the difference is not obvious from outside. And a straightforward not yet: if you have no rituals or agreements in place at all, the ACT skills have nothing to attach to. Build the two-minute check-in first, run it for a fortnight on easy days, and add the harder material once there is a container for it.

Where this does not apply

If the relationship is not safe in a basic physical or emotional sense, PACT's founding assumption — that both partners are protecting a shared unit in good faith — does not hold, and no amount of acceptance work should be used to talk anyone into tolerating harm. In that situation the two practices are not enhancing each other. Recommending acceptance there reads as advice to make peace with a bad situation rather than leave it, which is not what either model is for, and an unsafe partner will use the vocabulary faster than you will. Second, there is a way this pairing produces the opposite of what it promises. A couple in the early, high-conflict stage of learning PACT's rituals may find that adding an introspective vocabulary on top hands an avoidant partner a more sophisticated way to intellectualise instead of showing up — I'm noticing some defusion around that is a very effective way of not answering a question. The honest fix there is more structure, not more inner work: fewer concepts, shorter rituals, fixed times. Third, and this one is a category error worth stating plainly, because it is common and it does damage: acceptance in ACT is for your own internal experience, not for your partner's behaviour. I am accepting that you are always late is not acceptance, it is a complaint in fancy dress, and it will be heard as one. Your feelings about the lateness are yours to hold without struggling; the lateness itself is a legitimate subject for an agreement, and folding it into acceptance language removes it from the negotiating table where it belongs. Fourth, if one of you is in active treatment for something serious, do not let a couples ritual become the delivery mechanism for it. A partner is not a therapist, the roles conflict, and the check-in works precisely because its demands are small.

What the evidence does and does not support

The two sides carry real evidence and it is not the same evidence. Several of ACT's core processes have randomized-trial support, principally for anxiety and mood outcomes in individuals: allowing an internal experience rather than struggling with it, and stepping back from a thought to watch it as a passing event rather than living inside it, are both supported that way. Behavioural activation — acting before motivation shows up, which is the plank under the do-the-ritual-anyway instruction here — has its own substantial trial base in the treatment of depression, and it is arguably the best-evidenced single element on this page. Values clarification is on softer ground, resting on clinical rationale rather than isolated trials. PACT's claims — that attunement needs structured attention rather than goodwill, and that an explicit mutual commitment changes how partners weigh ordinary conflicts — come from Tatkin's clinical model and case-based writing, not from randomized comparison, and the claim about rituals is the weakest of the three, supported anecdotally. Note the mismatch that matters most: almost all of ACT's trial evidence is for individuals with a clinical presentation, not for couples and not for the behaviour of a partner under conflict, so importing it here is an extrapolation across both population and outcome. Nothing has tested this combination. A study that would move the question would randomize couples already using structured agreements to add a brief ACT skills component, and would measure something behavioural rather than a satisfaction score — whether the agreed ritual actually occurred on the nights both partners rated as bad. That measurement is the whole claim, and until somebody makes it, what you have here is an argument from mechanism with a well-evidenced neighbour, Integrative Behavioral Couple Therapy, standing next to it.