Trauma bonding is not love

Trauma bonding names a specific, documented dynamic, not love that has gone wrong and not any attachment that is simply hard to leave. It emerges from two structural conditions present together: a power imbalance, and intermittent reinforcement, abuse alternating with repair and affection, which produces a bond stronger than the relationship would otherwise form. The pull, the difficulty leaving, and the returning are produced by the structure, not by weakness, and the same conditions that build the bond are the ones that tend to escalate the harm.

Written and reviewed by Natalia Cajide, Registered Psychologist (AHPRA PSY0002474316). 6 min read. Published 7 June 2026. Last reviewed 7 June 2026. Next review 7 June 2027.

Need help right now? Crisis 000 · Lifeline 13 11 14 · Beyond Blue 1300 22 4636 · Suicide Call Back 1300 659 467

Where the term comes from

Trauma bonding is one of the phrases that gets used in two very different ways, and the slippage matters. In therapy rooms, the term names a specific dynamic with documented features. On social media, the term has become a way of describing any difficult attachment, including healthy attachments that are simply hard to leave. The two uses are not the same thing, and one of the more careful pieces of work I do, when someone arrives wondering whether they are in a trauma bond, is to slow down and look at what is actually there.

What the academic literature calls trauma bonding is not love that has gone wrong. It is a particular dynamic that emerges in specific structural conditions, with a recognisable pattern that has been documented in research for forty years. Understanding the pattern is part of how a person can recognise what they are in.

The concept of traumatic bonding was first articulated by the psychologists Don Dutton and Susan Painter in 1981. Working with women who had stayed in or returned to relationships with men who had been violent toward them, Dutton and Painter observed two specific structural features that the women's accounts had in common. The first was a power imbalance in the relationship. The second was the intermittent nature of the abuse: periods of frightening or violent behaviour alternating with periods of remorse, affection, and apparent change.

Their argument, supported by subsequent research, was that this combination produced a particularly strong attachment, stronger than what the same person might have formed in a relationship without those structural features. The bond was not evidence of love or weakness. It was the predictable outcome of the structural conditions. Other research has since used the term traumatic bonding for similar dynamics in other relationships of intermittent abuse, including sex trafficking situations and high-control groups (Casassa et al., 2021).

The structural conditions

The dynamic that produces a trauma bond has two features that need to be present together.

The first is a power imbalance. One person in the relationship has substantially more control than the other. The control can be physical, economic, immigration-related, social, or a combination. It is not the absence of all power that creates the bond. It is the structural asymmetry.

The second is intermittent reinforcement. The abusive behaviour is not constant. It is unpredictable, alternating with periods of repair, affection, remorse, or apparent commitment to change. The unpredictability is not incidental. The intermittent pattern is itself part of what makes the bond strong, because the system that learns associations cannot easily decouple the abuser from the relief that follows the abuse.

These two features, together, produce what looks like and feels like love but is structurally something else. The pull toward the relationship, the difficulty of leaving, the return after leaving, the inability to see the relationship clearly from inside it are all patterns produced by the structure. They are not patterns produced by personal weakness.

Why "Stockholm syndrome" is the worse framing

You will sometimes hear this dynamic called Stockholm syndrome, after the 1973 Stockholm bank robbery in which the hostages developed apparently positive feelings toward the bank robbers. The term entered popular use and is still common in cultural discussion.

In the academic literature, Stockholm syndrome is poorly supported. It is not in the Diagnostic and Statistical Manual or the International Classification of Diseases. The original 1973 case has been re-examined and the evidence for the syndrome as originally described is weak. Researchers in the field of intimate partner violence have largely moved away from the term in favour of the better-supported concepts of traumatic bonding and coercive control (Casassa et al., 2021).

The coercive control framework developed by Evan Stark in 2007 names what is happening in many of these relationships. Coercive control is a strategic pattern of behaviour by which one person systematically undermines the autonomy, decision-making, and resources of another (Stark, 2007). It is not the same as a series of bad arguments or even a series of violent incidents. It is a system. In Australia, coercive control is recognised in research and policy as a distinct form of harm, and several Australian jurisdictions have moved to criminalise it (Australia's National Research Organisation for Women's Safety, 2021).

What the Australian data shows

The Australian Institute of Criminology's 2021 study of coercive control among Australian women (Boxall & Morgan, 2021) surveyed 1,023 women who had recently experienced coercive control. The findings underscore how often coercive control is not the absence of physical harm but a precursor and accompaniment to it.

Over half of the surveyed women (54 percent) reported experiencing physical forms of abuse alongside the coercive control, including severe forms. Approximately one in four (27 percent) reported non-fatal strangulation, which is a known risk factor for subsequent intimate partner homicide in the international research. Approximately one in three of the surveyed women (30 percent) reported experiencing sexual violence during the survey period.

These figures matter in the trauma bonding conversation for one reason. The structural conditions that produce a trauma bond, the power imbalance and the intermittent reinforcement, are statistically not separate from the conditions that produce physical and sexual harm. They are the same conditions. When the dynamic of trauma bonding is present, the risk of further harm is also present.

What stays unresolved

There is no version of this conversation that does not sit with two true things at once. The first is that what you may be feeling, including the love, the care, and the genuine hope, is not artificial. The intermittent reinforcement does not produce a fake attachment. It produces a real one, that has met a real need at moments, and that is part of what makes leaving harder than it ought to be. The second is that the structural conditions producing the attachment are also the conditions producing the harm, and the research suggests they tend to escalate rather than settle. Both of those things are true. Acknowledging the first does not soften the second, and acknowledging the second does not dismiss the first.

What I would offer in place of a recommendation is the framing of the right question. The right question is not whether to love the person less. It is what the structural conditions are, and what would be safer to do given those conditions, taking the full picture into account. That conversation belongs with someone who can hold all of it at once, including the affection and the danger, the hope and the cost, and the practical realities you are inside. The footer of this page is the start of finding that person.

This content is general information only. It is not a substitute for individual psychological or medical advice. Reading this does not establish a therapeutic relationship with Equal Psychology or any of their clinicians.