Power in the Couple

In every relationship, there is someone who decides more often. Someone whose mood weighs more heavily on the evening. Someone who can withdraw and make the other wait. This is not a design flaw: it is the ordinary mechanics of two people sharing a life.

The problem is never the existence of power. It is its immobility.

This chapter offers a vocabulary for naming what, in a couple, almost never gets said — who decides, who gives way, who has the final word on what it is legitimate to feel — and a method for rebalancing before resentment becomes structural.

First: imbalance is not control

This distinction governs everything that follows. Without it, what comes next becomes a machine for reframing any annoyance as abuse — or, worse, for reframing abuse as a mere "imbalance to negotiate".

Every relationship carries asymmetries, and they reverse over time. An asymmetry is not abuse. Three criteria separate the two:

Systematicity. Is it an episode, or the permanent regime of the relationship? A partner who imposes their decision one tired evening is not a partner who has been imposing their decisions for four years.

Fear. Does one of the two durably adjust their behaviour to avoid the other's reaction? Choosing words carefully, anticipating moods, giving up on mentioning certain subjects: this is the most reliable sign, and the earliest detectable one.

Contraction. Does one person's freedom — relationships, money, movement, expression, work — shrink over time? Control is measured less by what it imposes than by what it takes away.

Without these three elements, we are talking about imbalance: that is the subject of this chapter, and it can be negotiated. With them, we are talking about control: that cannot be negotiated, and falls under Parts 2 and 4.

Two ways of holding power

Relational power is not a cake. What one takes, the other does not necessarily lose — and it is precisely the opposite belief that turns a couple into a battlefield.

Power-over is the capacity to impose one's will: decide, settle matters, obtain. It works — in the short term. Over time, it invariably produces the same thing: passive resistance, withdrawal, and the erosion of intimacy. One gains compliance and loses the person.

Power-with is the capacity to influence and be influenced. It is the power of the one who obtains because they have persuaded, and who agrees to be persuaded in return. It is slower and far sturdier.

The difference comes down to a single question: can my partner make me change my mind? If the answer is no — never, on any subject — you are not in a relationship of mutual influence, whatever the apparent comfort of the arrangement.

Gottman and Levenson's (1992) work on marital stability points in this direction: the systematic refusal to accept the other's influence is a bad omen, and contempt — the most accomplished form of power-over — is the most powerful predictor of breakup.

Power as responsibility, or as privilege

Same amount of power, two opposite uses.

As privilege: I carry more weight, so my needs come first, I owe no justification, and it is normal that things revolve around me.

As responsibility: I carry more weight, so I must pay attention to how I use it, check that the other has room to say no, and actively compensate for the asymmetry.

It is the same imbalance. It is not the same relationship.

The invisible forms

The most durable imbalances are not about money or big decisions. They are about things no one thinks to count.

The power to define reality

It is the capacity to say what really happened, and what it is legitimate to feel.

In a couple, the one with the richer emotional vocabulary often ends up arbitrating. This expertise is real and usually acquired at a cost — it comes from work, from therapy, or from a childhood where one had to learn to read others in order to survive. It is legitimate. But it can also close down discussion: when one of the two holds the lexicon, the other eventually gives up trying to describe their experience, unable to name it as precisely.

The right question is not who is right about what I feel. It is: do we each have a place to describe our experience without it being immediately corrected?

Moral power

It is the capacity to position oneself as the wronged party, and thereby secure support that is hard to contest.

It works by accumulation: I work for this family, I carry the entire mental load, I gave up my career. Each of these statements may be accurate. They become moral power when they are used to close a discussion rather than open one — when sacrifice is invoked as a debt owed.

Distinctive sign: the conversation can no longer be about the problem, only about the comparative merit of the two people.

The power of withdrawal

It is the capacity to threaten, explicitly or not, to withdraw — emotionally, materially, or altogether.

Its force lies not in the threat but in the uncertainty it installs: the partner who dreads the withdrawal spends a growing share of their energy preventing it. The prolonged silence after an argument is its most common form, and the most effective.

The power of expertise

It is the capacity that comes from exclusive mastery of a domain: the one who manages the finances, the one who knows the children's medical files, the one who understands how the household runs.

It is the most benign of the four — a normal division of labour — and the most insidious over time. It produces dependence without intent. The test is simple: if you disappeared for three weeks, what would collapse? What would collapse is exactly the measure of your power of expertise — and of the other's vulnerability.

The competition cycle

When power stops being debatable, couples enter a predictable sequence.

1. Perception of threat. One of the two judges that the other has more freedom, time, or weight in decisions.

2. Building the file. Each gathers their arguments and their evidence. Scores are kept. Things are remembered. Friends and family are mobilised as witnesses.

3. Tactical escalation. The means grow more refined: invalidation, irony, withdrawal, allusions to leaving, children enlisted as witnesses.

4. Exhaustion. Both feel less understood than at the start of the conflict, and less able to address the real problem — which was never discussed.

5. Entrenchment. Every interaction becomes a stake. The relationship is no longer fought over subjects: it is fought over itself.

The tipping point is step 2. As long as you are discussing a problem, you can solve it. From the moment you start building a file, the objective has changed: it is no longer about solving, but about being right.

Comparative table

Aspect	Settled imbalance	Living balance
Influence	One never changes their mind	Each can persuade the other
Decisions	Always the same reserved domains	Arbitration rotates according to the stakes
Justification	One must justify themselves, the other does not	Both explain their choices
Disagreement	Costs dearly, so it is avoided	Ordinary, with no consequence on the bond
Reversibility	The imbalance has frozen	It reverses across periods
Autonomy	"Asking permission"	Own time, money, relationships not negotiated
Accounting	Scores are kept	Forgetting to count
Repair	The conflict leaves a residue	The conflict is settled and closed
Use of power	As privilege	As responsibility

The invisible signs

What goes unnoticed, because it has become the normal texture of the relationship:

• You know in advance which subjects are not worth the fight. The list keeps growing.
• You announce your personal plans instead of proposing them — or the reverse: you propose them instead of announcing them.
• One of you has been saying "I'll let you handle it" for an entire domain, for years.
• Your friends are no longer really your friends, without any prohibition ever having been spoken.
• There are expenses you justify, and others you do not — and it is not correlated with the amount.
• After an argument, it is always the same person who comes back first.
• You cannot answer the question: what have I changed my mind about thanks to my partner, this year?

None of these signs is serious in isolation. Three or four together describe a relationship that has frozen.

The lucidity test

Six domains. Answer for each: who decides? — me / my partner / truly both.

1. Money. Who decides on major spending? Who holds a de facto veto? Who has to justify their personal purchases?
2. Children. Who settles questions of upbringing? Who is recognised as "the one who knows"? Who handles the unexpected?
3. Personal time. Who can step away without negotiating? Who takes over during that time?
4. Career. Who was able to prioritise theirs? Who arbitrated in favour of the family — and was it decided, or did it just happen?
5. Intimacy. Who initiates? How are divergent desires settled? Does a refusal have consequences for the mood of the following days?
6. Emotional life. Who is allowed to struggle? Who has to hold it together? Who comforts whom, and how often?

Method: each answers alone, in writing, before any discussion. Then you compare.

What to read in the results. The score does not matter — these are not points to be equalised. Two things matter:

• The gaps in perception. When one answers "both" and the other "you", the gap is the information. It signals an imbalance that is invisible to the one who benefits from it — that is the very definition of a relational privilege.
• Concentration. A distributed imbalance (one decides on three domains, the other on three) is a division of labour. The same imbalance concentrated on a single person is a hierarchy.

Recommended frequency: once a year, and systematically after a change in configuration — a birth, a move, a change of role, an illness.

Rebalancing

Do not turn the audit into a trial

The most common mistake: the partner who discovers an imbalance to their disadvantage arrives with their conclusions already written. The other defends themselves. Rebalancing becomes yet another conflict.

The object of the conversation is not who abused whom. It is: how did we get here, and what do we change? Most imbalances are wanted by no one. They settle in by default, by competence, by exhaustion.

Choose two domains, not six

Identify together two domains to rebalance, with a concrete action and a deadline for each. A general rebalancing always fails; a precise transfer holds.

And a transfer of power presupposes a transfer of competence: the one who takes back a domain must have the right to do it badly for a few months. The partner who takes back the management of a budget and gets corrected at every decision has taken back nothing at all.

Four rules of conversation

Rephrase before answering. "I hear you saying that…" before "Well, I…". It feels artificial the first few times. It is. It works anyway.

Validate before contesting. Find what is right in the other's point of view before laying out your disagreement. Validating is not giving in.

Speak in "I". "I feel overwhelmed" opens a discussion; "you do nothing" closes it.

Prefer curiosity to verdict. When the other's answer surprises you, the right reaction is "explain it to me", not "that's wrong".

These rules are not mere politeness. Under stress, the prefrontal cortex — the one that lets us reason and hold back — yields to faster, more defensive circuits (Arnsten, 2009). These four rules are speed bumps: they keep the conversation in a regime where we are still able to think.

The stop signal and repair

A word agreed in advance means: I am overwhelmed, we pick this up later. It is neither a confession nor a victory, and it must be honoured immediately — otherwise it is worth nothing. Set the resumption within twenty-four hours: without a deadline, the stop signal becomes a form of the power of withdrawal.

Conflict rules: no insults, no threats of breaking up, no recruiting allies mid-conflict, no list of past grievances, and a capped duration with a mandatory break.

Repair is the part everyone skips. After a significant conflict: say what you understood, what you regret about how it went, and explicitly close it. Without this gesture, each argument leaves a residue, and residues add up far faster than arguments.

When inequality is inevitable

Illness, depression, job loss, bereavement, the arrival of a child: certain periods make the relationship frankly unequal. This is normal, and denying it does more damage than the inequality itself.

An inequality is not a hierarchy. The partner who temporarily needs more loses neither their dignity, nor their voice, nor their right to disagree.

Reciprocity is temporal. What matters is not the balance of the month, but that the roles have rotated over time. A couple where support has never changed direction in ten years is not going through a rough patch: it is inside a structure.

The supporter's limits are legitimate. The one who carries must be able to say they are tired, without it counting as abandonment. A carer who burns out in silence ends up resenting the one they help — and no one saw it coming.

Name the way out. When the period ends, take the time to say so, to thank explicitly, and to redefine the roles. Otherwise the emergency configuration becomes the default configuration, and no one remembers why.

Convergence of wisdoms

Al-Ghazâlî places Justice among the four cardinal virtues, and defines it as balance: each thing in its place, each faculty in its measure, neither in excess nor in deficit. Power-as-privilege is an excess; the self-effacement of the one who never asks for anything is its symmetrical deficit. Both break the same virtue — and this is a useful intuition, because we spontaneously see only the first.

The Stoics, with Hierocles' concentric circles, describe concern for others as a progressive widening: from the self to those close, then to humanity. Power-with is exactly this movement applied to the couple — considering the other's interests as entering one's own, not by sacrifice, but because the circle has widened.

Jung describes the power complex as a compensation: one dominates where one feels threatened. This does not excuse it, but it explains why imbalances resist argument so well — one does not reason one's way out of what protects against a fear one will not admit.

Exercises

Exercise 1 — Three weeks

Note down, for twenty-one days, one line each evening: one decision made that day, and by whom. Analyse nothing before the end. Reread at the close of the three weeks. The distribution is almost always sharper than either believed — in one direction or the other.

Exercise 2 — The cross-audit

The lucidity test above, filled out separately, then compared. One hour, once a year. Aim: identify the gaps in perception, then choose two domains to rebalance with a dated action for each.

Exercise 3 — The yearly question

A single question, spoken aloud once a year:

What have I changed my mind about thanks to you this year?

Each answers. If one of the two can find nothing, it is not a conversational detail: it is the fastest diagnosis there is of a relationship where influence no longer circulates in both directions.

Exercise 4 — The expertise inventory

Each lists the domains they alone master, and what would happen if they disappeared for three weeks. Then choose one to actually transfer: documentation, accompaniment, then complete withdrawal. The transfer succeeds the day you are no longer consulted.

Summary

1. Power is not the problem; its immobility is. Every relationship contains some, and it must circulate.
2. Three criteria separate imbalance from control: systematicity, fear, contraction of freedoms. Without all three, you negotiate; with all three, you do not.
3. Power-over versus power-with. The test comes down to one question: can my partner make me change my mind?
4. Four invisible forms: defining reality, moral power, withdrawal, expertise. These are what structure the relationship, not the big decisions.
5. The competition cycle tips at the file-building step — the moment one stops wanting to solve and starts wanting to be right.
6. The gap in perception is the information. An imbalance invisible to the one who benefits from it is the definition of a relational privilege.
7. A transfer of power presupposes a transfer of competence, and the right to do it badly for a few months.
8. Temporary inequality is normal; permanent inequality is a structure. What matters is that the roles have rotated over time.
9. Repair after conflict is not optional. Without it, residues add up faster than arguments.