From the outside, harmful relationships can look strangely simple.
Someone is being mistreated. They know they are being mistreated. So why not leave?
Why keep defending the person who hurts you? Why miss them after you escape? Why return? Why protect an institution that failed you? Why become angry at the person who points out the harm?
Those questions make sense if we imagine attachment and harm as opposites.
Human relationships are rarely that clean.
The person who frightens you may also be the person who comforts you. The institution that silences you may also contain your friends, your identity, your history, and your sense of purpose. The relationship that restricts your freedom may still contain affection, good memories, hope, and periods when everything feels normal again.
Sometimes loyalty survives harm because the harm and the attachment have become tangled together.
TL;DR
People can remain attached to harmful relationships for reasons that go far beyond simply “choosing” abuse. Power, dependency, isolation, fear, affection, hope, economic constraints, and intermittent periods of kindness can all shape what feels possible. Research on traumatic bonding specifically examines attachment under conditions of abuse, power imbalance, and intermittent treatment; it should not become a casual label for every dysfunctional relationship or intense loyalty. In institutions and high-control groups, concepts such as institutional betrayal, coercive loyalty, dependency, and bounded choice are often more accurate. Understanding why loyalty persists does not excuse the person causing harm. It helps us see how attachment itself can become part of the structure that makes leaving difficult.
When the Person Hurting You Is Also the Person You Need
Jennifer Freyd’s betrayal trauma theory begins with a difficult problem: what happens when the person you depend on is also the source of danger?
A child dependent on an abusive caregiver cannot simply evaluate the relationship the way an independent adult might evaluate a bad friendship. Maintaining some connection to the caregiver may be necessary for food, shelter, protection, affection, and basic survival. Freyd argues that under conditions of dependency, fully recognizing betrayal can itself become dangerous because recognition normally pushes us toward confrontation or withdrawal, options the dependent person may not actually have.
Something similar can happen in adulthood without requiring exactly the same mechanism. A partner may control money, transportation, housing, friendships, or access to children. They may alternate intimidation with tenderness. They may apologize convincingly. Months of tension may be followed by a weekend that reminds the person why they fell in love.
Leaving is no longer one decision made in one moment.
It is a decision being made inside an entire relationship.
When Kindness Becomes Part of the Trap
This is where the term traumatic bonding has a legitimate, narrower meaning.
Donald Dutton and Susan Painter studied women who had recently left abusive intimate relationships. Their 1993 research found that stronger power imbalances and more intermittent maltreatment were associated with greater continuing attachment to former abusive partners. Even six months after leaving, characteristics of the abusive relationship predicted a substantial amount of the variation in how strongly participants remained attached.
The intermittency matters.
If someone were cruel every minute of every day, the relationship might be easier to understand.
Instead, harm can be mixed with ordinary life. There are birthdays and jokes and sex and apologies and promises. There may be genuine affection alongside control.
The good moments do not erase the abuse.
But the abuse does not magically erase every attachment that existed before it either.
That is why “Why do they still love them?” is sometimes the wrong question.
Love may be part of what makes the situation difficult.
Control Is Bigger Than Violence
Evan Stark pushed domestic-violence research toward another important distinction: sometimes the central problem is not a series of violent incidents but a pattern of coercive control.
A person’s life can gradually become smaller. Their contacts are monitored. Their money is restricted. Their decisions are questioned. Privacy disappears. Everyday choices begin carrying consequences. Stark describes coercive control as an assault on autonomy and liberty, not merely a collection of assaults.
Modern evidence supports treating this seriously. A 2024 systematic review and meta-analysis covering dozens of studies found moderate associations between exposure to coercive control and both PTSD and depression.
This helps correct another common misunderstanding.
Someone may not stay because they have been psychologically hypnotized into loving abuse.
They may be entrapped.
Their choices have been narrowed over time.
Judith Herman describes trauma around the experiences of disempowerment and disconnection. Recovery therefore requires more than convincing someone that the relationship is bad. It requires restoring safety, power, connection, and the ability to make meaningful choices again.
Institutions Can Become Part of the Attachment Too
Relationships are not the only places where loyalty and harm coexist.
People depend on churches, universities, workplaces, militaries, political organizations, activist groups, and other institutions for community, identity, status, livelihood, and meaning.
Jennifer Freyd uses the term institutional betrayal for situations where institutions harm people who depend on them, including by failing to prevent abuse or responding badly when abuse is reported.
That creates an uncomfortable conflict.
Recognizing the betrayal may mean admitting that people you trusted failed you. Speaking publicly may threaten friendships, employment, family relationships, or an identity built over decades.
Sometimes defending the institution protects more than the institution.
It protects the person’s entire understanding of where they belong.
When Choice Becomes Bounded
Sociologist Janja Lalich offers another useful concept from her research on charismatic and high-control groups: bounded choice.
People inside such groups are not necessarily irrational, unintelligent, or incapable of making decisions. Lalich argues that a tightly organized social environment can gradually shape the assumptions within which choices are made. Decisions that look incomprehensible from outside can make sense from inside the system because the system has helped define what loyalty, morality, danger, and betrayal mean.
This gives us an important boundary for the language we use.
A devoted Catholic is not “trauma bonded” to Catholicism because they strongly defend their church.
A committed socialist is not trauma bonded to a political organization because leaving would hurt.
Strong belonging is not automatically pathological.
But when an organization combines dependency with isolation, punishment for dissent, manipulation, control of information, or retaliation against people who leave, ordinary loyalty can become something more coercive.
Calling that high-control attachment or coercive loyalty may tell us much more than casually calling it a trauma bond.
Loyalty Is Not Proof That Harm Did Not Happen
One of the cruelest consequences of these dynamics is that continued attachment can be used against the harmed person.
If it was really abuse, why did you stay?
Why did you go back?
Why did you defend them?
Why do you still miss them?
But those questions assume that harm should instantly destroy attachment.
Sometimes the opposite happens. The relationship becomes more psychologically complicated precisely because safety, fear, affection, dependency, identity, and hope all occupy the same space.
Understanding that complexity is not about removing agency.
It is about understanding the conditions under which agency is being exercised.
And that changes what help looks like.
If harm works by reducing someone’s world, recovery cannot simply mean telling them what to do next. It means helping the world become larger again.
More safety. More relationships. More information. More material options. More room to disagree. More ability to imagine a life beyond the system that once seemed impossible to leave.
The goal is not to replace one person’s control with someone else’s.
It is to give choice back.
Practice for the Week
When you see someone remaining loyal to a person or institution that appears to be hurting them, notice the impulse to ask:
Why don’t they just leave?
Then try a different question:
What might leaving cost them, and what would have to exist for leaving to become a real choice?
Influencers
Judith Lewis Herman — Trauma, captivity, disempowerment, disconnection, and recovery through restored safety and agency.
Jennifer Freyd — Betrayal trauma, betrayal blindness, institutional betrayal, and the consequences of harm by people or systems upon which someone depends.
Evan Stark — Coercive control as a pattern of domination and restriction of autonomy rather than merely isolated acts of violence.
Donald Dutton and Susan Painter — Early empirical research examining traumatic bonding, intermittent maltreatment, power imbalance, and continuing attachment in abusive intimate relationships.
Janja Lalich — “Bounded choice” and the social dynamics through which high-control groups can narrow what choices seem reasonable or morally available to members.
Next week: When a Society Cannot Remember Honestly.