Using Secure Functioning to enhance Token Economy

What each one is actually for

Secure Functioning

Secure functioning starts from an unsentimental premise: goodwill decays. Two people who love each other and intend well will still drift apart under ordinary conditions — fatigue, work, children, the slow accumulation of unrepaired small moments — because nothing in the arrangement actively resists the drift. Tatkin's answer is to make the protection explicit rather than assumed. Partners agree, out loud and in specific terms, that they will put the relationship's security ahead of being right, ahead of their own comfort in a given argument, ahead of the various other loyalties that pull at it. He calls the resulting arrangement a couple bubble, meaning a mutual agreement that each person's job is to look after the other's nervous system and to be the one they can rely on first. Then the model gets concrete in a way that surprises people expecting therapy to be about feelings. It asks for specified rituals around the moments where couples reliably fail: how you say goodbye in the morning, how you greet each other at the door, what happens at bedtime, what you do in the first sixty seconds after a fight. Those moments are chosen because they are frequent and because leaving them to whoever happens to have the energy means they gradually stop happening.

Token Economy

A token economy is one of the plainest behavioural systems ever built. You name the target behaviours precisely enough that two people watching would agree on whether one occurred. Someone earns a token each time they do one — a chip, a point, a mark on a chart. Tokens are exchanged later for something the person genuinely wants and chooses themselves. The reason for the token rather than the reward directly is timing: reinforcement works best delivered immediately, and the things people actually want usually cannot be handed over on the spot, so the token stands in as a placeholder that keeps the connection between the action and the payoff intact. The approach was developed in psychiatric hospitals in the 1960s and now appears most often as contingency management in substance use treatment, where a person receives vouchers or prize draws for drug-negative tests. The reason it is taken seriously is that it works better than almost anything else in that setting. The reason it is contested is a finding that shows up repeatedly: gains often shrink once the tokens stop, which suggests the system holds the behaviour up rather than installing it.

What you get out of pairing them

What secure functioning offers is protection against a failure mode that is almost invisible while it is happening. Nobody notices connection eroding, because it does not erode through events — it erodes through the absence of them, through six months of goodbyes shouted from the hall and evenings on separate screens, none of which registers as a problem on any given day. The model's usefulness is that it makes the small repeated moments explicit and therefore checkable, which turns a drift you cannot see into a set of agreements you can name.

What a token economy offers, at its best, is not motivation. It is precision. The system cannot function unless the target behaviour is specified so exactly that two observers would agree on whether it happened, and that requirement forces a clarity almost no other approach demands. 'Be more supportive' cannot be counted, so it cannot enter the system. 'Ask me one question about my day before eight o'clock' can. That constraint, adopted on its own, exposes how much of what couples agree to is unmeasurable and therefore unfalsifiable.

What the pairing gives you is one narrow, honest thing. Secure functioning asks for explicit agreements and rituals but supplies no mechanism for knowing whether they are actually happening, and both partners' memory of the last fortnight is coloured by mood — a resentful person and a contented person will report the same two weeks differently and both will be sincere. A shared, visible record of a precisely named agreement settles that. Not a scoreboard, not a ledger, and emphatically not something anybody trades against. A record. That distinction is the whole page, and getting it wrong turns a useful borrowing into the thing the research warns about.

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 Token Economy rests on

• Small wins compound randomised trials
  Repeating a small consistent action accumulates over time into a result too large to produce in a single step.
• Contingent reinforcement strengthens behavior randomised trials
  A behavior followed reliably by a valued reward increases in frequency over time.
• Secondary reinforcement bridges delay plausible mechanism
  A symbolic token standing in for a later reward keeps reinforcement effective when the real reward cannot be delivered immediately.

Where they actually meet

The two agree about something fundamental and disagree about something equally fundamental, and both halves are worth having in view. The agreement: neither trusts intention. Tatkin's position is that two people can love each other sincerely and still let the relationship degrade, because sincerity does not schedule anything, which is why he asks for named agreements and repeated rituals rather than for better feelings. Behavioural work reached the same conclusion from the opposite end of psychology — that what a person intends predicts their behaviour considerably less well than what reliably follows it — and built systems that ignore intention entirely. Two traditions with almost nothing else in common both concluded that structure outperforms goodwill, which is at least suggestive. The disagreement is about what may be made conditional. A token economy works by making something desirable contingent on performance. Secure functioning works by making protection explicitly non-contingent, on the reasoning that a partner who has to earn safety is not safe, and that the vigilance required to track whether you are ahead or behind is itself the thing that damages the bond. That is a real conflict and not a matter of emphasis, which is why the borrowing here has to be narrow.

The principle underneath both

Both hold that what you want to happen reliably has to be built into structure rather than left to whoever feels like it that day. Tatkin arrives there through clinical observation of couples: the ones who stay connected are not the ones with more affection available, they are the ones who have made the frequent small moments non-optional, so that a bad Tuesday costs one goodbye rather than starting a slide. Behavioural work arrives there through decades of data showing that a reliably delivered consequence changes what people do far more dependably than a resolution to change does. So both are, in their different vocabularies, distrustful of the moment and respectful of the arrangement made in advance. Where they part company is on what belongs in the arrangement — and it is worth being explicit that this shared framing is a reading offered here rather than something either tradition would claim. A behaviourist would not describe a couple bubble as a reinforcement schedule, and Tatkin would object, reasonably, that his agreements are about safety rather than about behaviour frequency.

What that makes the real goal

Name the goal without either tradition's language and it comes out as this: making the thing you actually want to happen happen reliably, without requiring anybody to feel like it on the day. Once stated that way it is separable from couple therapy and from behavioural treatment alike, and it produces three questions you can put to any agreement you have made with another person. Is the behaviour specified precisely enough that both of you would agree afterwards on whether it happened? Is there any way of knowing whether it is occurring that does not depend on either person's memory, which is mood-coloured in both directions? And is anything about the other person's security made conditional on it — because if it is, you have built the version that damages the relationship, whatever the intention was. Almost every agreement couples make fails the first two and passes the third by accident.

Other practices that serve the same goal

Several practices aim at the same target and carry evidence of very different weights. Behavioural couples therapy is the most directly relevant, and its history is the single most useful thing on this page. The field did try contingency contracting between partners, and it distinguished two forms: quid pro quo contracts, where each partner's change is explicitly conditional on the other's, and good faith contracts, where each commits to change independently. The general finding across that era of work was that the good faith form held up better and the quid pro quo form was fragile — the moment one partner slipped, the other's obligation lapsed by design, and the arrangement collapsed into scorekeeping. That is close to a direct test of the danger this page is warning about.

John Gottman's rituals of connection serve the same objective as Tatkin's and come with a much larger observational research base — decades of longitudinal work on which couples stayed together, with findings about repair attempts and the ratio of positive to negative interaction during conflict. It is predictive rather than experimental research, but there is a lot of it.

Implementation intentions offer the leanest version: a written 'if this happens, then I will do that' sentence attached to a specific cue. That literature has extensive randomised support, mostly for individual health and achievement behaviours rather than relational ones, and the transfer to a couple's context is an assumption rather than a finding.

And habit stacking, attaching the new action to something you already do reliably every day, solves the practical problem all of these share: without a cue that fires whether or not anybody remembers, agreements decay. Its own evidence base is thin and borrowed, but the cue problem it addresses is real and unglamorous.

A tradition that already does this

The precedent here is a cautionary one, and it is more valuable than an encouraging one would be. Behavioural marital therapy in the 1970s and early 1980s did attempt to bring exchange structures into intimate relationships — partners negotiated explicit contracts specifying who would do what, sometimes with agreed consequences attached. The field largely moved away from the exchange-based version, and the reason it gave was not squeamishness. Quid pro quo arrangements were observed to be brittle in exactly the way their design predicts: they require continuous monitoring of whether the other person has held up their end, that monitoring is corrosive to the trust the couple came in to repair, and a single defection dissolves the obligation on both sides. What the field kept was the part this page is recommending — behavioural specificity, agreements stated concretely enough to be checked, and attention to what actually happens rather than to what each partner believes happens. So the narrow borrowing proposed here is roughly the borrowing that survived fifty years of clinical experience, and the broad one is roughly what was tried and abandoned.

Using an unconditional commitment to fix what token systems are worst at

The persistent finding about contingency systems is that behaviour often falls back once the tokens stop. This is not usually treated as a mystery — the arrangement was holding the behaviour up, and when the support goes, so does the behaviour. The interesting question is what makes an action persist after the external structure ends, and the most common answer, both in the research on maintenance and in what people say about their own change, is that it became attached to something they cared about beyond the reward. A relationship they did not want to disappoint. An identity they had started to hold.

Secure functioning is, among other things, a method for making that attachment explicit rather than hoping it forms. That has a direct practical application for anyone running an actual token system with a child, which is the setting where couples most often encounter one.

Here is the concrete version. If you are running a chart with a child, run the agreement about the chart the way the couple model runs agreements with each other: state out loud, more than once, and particularly on the days it is going badly, that the chart is about a specific behaviour and that nothing about your relationship with them is on it. 'This is about the homework at six. You and me are not on this chart.' Say it after a day where they earned nothing, because that is the day it is genuinely in doubt from their side and the day it does the most work.

Then build in one thing token systems do not include: a fixed, unconditional point of connection every day that is entirely outside the system and cannot be lost. Ten minutes at bedtime, or whatever fits, that happens identically whether the chart had a good day or a terrible one. Its function is structural rather than sentimental — it is what keeps the child's security from becoming one of the things at stake in the arrangement.

Between two adults the same logic applies more sharply and points to a firmer conclusion. If either of you would feel your standing in the relationship rise or fall with the tally, the arrangement has become the thing behavioural couples work moved away from, and you should stop it rather than adjust it.

Borrowing behavioural precision to find out whether your agreement is actually being kept

Couples make agreements that cannot be checked, and then argue about whether they are being kept, which is an unresolvable argument because there is no fact to appeal to. 'We agreed you'd be more present in the evenings.' 'I have been more present.' 'You have not.' Both of them are reporting honestly and both are reading the same fortnight through their current mood, which is exactly what memory does under strain.

The borrowing is one requirement, and it is the requirement that makes a token system function at all: the behaviour has to be named precisely enough that two observers would agree on whether it occurred. Run your existing agreements through that test and most will not survive it. 'More present' fails. 'Phone in the drawer between the time we sit down to eat and the time the plates are cleared' passes. 'Handle bedtime more' fails. 'I do Tuesday and Thursday bedtime start to finish' passes. The exercise is uncomfortable, because vagueness is often doing something for both people — it lets each of you believe the agreement means what you wanted it to mean, and it protects you from finding out.

Second, borrow the visible record and nothing else. A shared note, a mark on the kitchen calendar, whatever is easiest. Three rules, and they are what keep this from becoming the thing this page has been warning about. Nothing is exchanged for it, ever — no favours, no leverage, no 'I've done four and you've done one.' It records only what happened, never a judgement about it. And either person can mark it, which prevents it becoming one partner's surveillance of the other, which is the most common way a good idea here goes bad.

What this is actually for is the conversation four weeks later, and what it changes is what that conversation has to work with. Instead of two people arguing about a disputed fortnight, you are both looking at a note that says the phone went in the drawer on nine of the last fourteen evenings. Nine is a genuinely different conversation from either 'I've been doing it' or 'you never do it.' It is a starting point instead of a standoff, and neither of you has to be wrong for it to be true.

Sunday night, six weeks after you agreed he would do the bedtime routine more

You are loading the dishwasher and it comes out flatter than you meant it to: 'You said you'd take bedtime more and nothing's changed.' He straightens up and you can see the whole conversation arrive on his face before he says anything, because you have both had this one before and neither of you enjoys it. 'That's not true, I've been doing it loads.'

And here is the thing — he is not lying. He is thinking of the three nights he did it while you were out, which cost him a lot and which he remembers vividly. You are thinking of the eleven nights you did it, which cost you a lot and which you remember vividly. Neither of you can produce the actual number, and the argument you are about to have is a proxy for the one you cannot have, which is about whether the other person notices what you do.

So you stop, which is hard, and you propose something small enough that it does not sound like an accusation: 'Can we just write it down? Not to keep score — I genuinely don't know what the number is and neither do you.'

What you agree takes four minutes. The behaviour, specified so there is nothing to interpret: whoever starts bedtime — bath through to lights out, no handover halfway — puts their initial in the box on the kitchen calendar. That is all. No target. Nothing owed. Nothing earned. Either of you can write it, including writing the other person's initial, which turns out to matter more than you expected because it makes the thing feel less like one of you inspecting the other and more like a shared note.

Then the part that comes from the other side of this page, and you do it deliberately: before it starts, he says something that sounds slightly formal out loud in a kitchen and is worth saying anyway. 'This isn't about whether I'm doing enough as a father. That's not up for review here.' You say the equivalent back. It takes eleven seconds and it is what prevents the calendar from becoming a document about his adequacy.

Three weeks later the boxes read eight of you, thirteen of him. He had been doing more than you thought and less than he thought, and you are both slightly wrong-footed, which is the useful outcome. The conversation that follows is about Wednesdays specifically, because that is where the pattern is visible, and it lasts about six minutes rather than an evening.

When this is worth reaching for

The same argument keeps happening — whether an agreement is being honoured — and neither of you can produce a fact, only a memory. That's the specific condition, and it's common: the dispute is unresolvable because the evidence on both sides is a feeling dressed up as a recollection.

An agreement made in good faith has quietly stopped happening, and neither of you decided to stop it. These things decay silently, and the absence of a record is what lets the decay go unnoticed until it surfaces as resentment about something else entirely.

Or a token system is already running somewhere in the household — most often with a child — and there's a known risk worth heading off: that the chart becomes the relationship. The unconditional daily connection point that sits entirely outside it is a small addition with a clear purpose.

None of this is needed when things are already going well: a couple who are connected and who both feel their contribution is seen doesn't need a calendar to prove it, and introducing one has a cost — it draws attention to counting in a relationship where counting wasn't happening. That cost is real and easy to overlook when you're enthusiastic about a method.

Where this does not apply

Do not attach any exchange to it. This is the line the page exists to draw, and it is not a matter of preference — a system where one partner's effort earns them something from the other is the arrangement behavioural couples work tried and moved away from, for the reason that it requires each person to track whether they are ahead or behind, and that tracking corrodes the trust the arrangement was meant to support. If a tally is being used as leverage in an argument, it has already become that.

Do not use a record where the imbalance is not in dispute. If one person is doing overwhelmingly more and both of you know it, measuring it more precisely does not address anything and can read as a way of avoiding the conversation by turning it into an administrative matter. In that situation the honest move is the conversation.

In a relationship involving coercive control, a record of what one partner has and has not done is a tool that can be used against them, and this whole approach is inappropriate there. That situation calls for specialist support, not a shared method.

And be careful with what the counting does to why you act. There is a real, though contested, line of research suggesting that attaching an external marker to something a person was doing willingly can reduce their appetite for it. The evidence for this is stronger for tangible rewards than for a plain record, and a mark on a calendar with nothing exchanged for it is a long way from a reward. Still, if either of you notices that the bedtime is now being done for the calendar, that is the point to stop — and stopping it is a small loss, which is a good reason to keep the whole thing lightweight enough to abandon.

What the evidence does and does not support

The gap between these two on evidence is unusually wide and it should not be smoothed over. Token economies and their modern form, contingency management, are among the better-supported behavioural interventions in existence — multiple randomised trials and meta-analyses in stimulant and opioid use disorder show effects that compare favourably to most alternatives in that field, and the approach has held up across decades and settings. It also comes with a consistently reported limitation: effects frequently attenuate after the contingency is withdrawn, which is a finding about maintenance rather than about whether the method works while running.

Secure functioning is on entirely different ground. PACT is a clinician's model built from attachment research, neuroscience and extensive case work, and it has an internally coherent account of why it should work. What it does not have is a body of randomised outcome trials comparing it to other couple therapies. That is not a criticism of Tatkin specifically — much of couple therapy sits in a similar position — but describing it as evidence-based in the sense that contingency management is evidence-based would misrepresent it.

The nearest thing to a test of any part of this pairing is old: the behavioural marital therapy work comparing quid pro quo contracts against good faith contracts, which found the exchange-based form more fragile. That is genuinely relevant to the warning here, though it comes from a different era, a different construct and small samples by current standards.

On the specific proposal — a precisely named agreement with a shared visible record and no exchange attached — nothing has been tested. A study could be described easily enough: randomise couples with a disputed division of labour into a specified-and-recorded condition, a specified-only condition, and usual practice, then measure relationship satisfaction and the actual distribution of the task at three months. Nobody has run it, and the case here rests on one well-supported ingredient, one clinically-derived model, and a historical warning about the version to avoid.