If you are not alone right now, this page will be in your browser history. You can read it in a private window, or clear it when you are done.
Difficulty trusting a new partner after an abusive relationship is not a damaged capacity for trust. It is a threat-detection system that was recalibrated by years of accurate information, and that has not yet been given enough evidence in a new setting to recalibrate back. The caution was correct in the place where it was learned. What it is now is out of date, which is a different problem from being broken.
Whether trust returns, how far, and how long that takes are not things anyone can promise, and this page does not. What it can do is explain the mechanism and separate two things that are constantly confused: a threat system that has learned something, and a person who has something wrong with her. Nothing here is an assessment of you or of anyone you are with.
Was I wrong to stop trusting?
You were not wrong. That has to be said before anything else on this page, because almost everything written about trust after abuse begins by treating the caution as the injury. In a longitudinal study of 246 women seeking help for intimate partner violence, who rated how likely they thought physical re-abuse was within the coming year and were followed up eighteen months later, the women’s own predictions “were more likely to be right than wrong, and were subject to neither a pessimistic nor optimistic bias” (Cattaneo, Bell, Goodman & Dutton, 2007).
Cattaneo’s study covers one population — women already in contact with services — and is not a general claim about human intuition. What it establishes is that what you were doing was a real skill producing real information. Women in this situation are usually described as poor judges of danger. When somebody actually tested it, they were not.
The skill was expensive to acquire. Reading a face for what is coming next, holding a sentence back until the room is right, noticing that a car parked differently — that is the trained output of living inside what coercive control actually is, where a missed signal was paid for the same evening. You did not develop a suspicion problem. You developed a detection system under conditions that required one. So the question is not what is wrong with me, but: what happens to a finely tuned detection system in a room where there is nothing to detect?
Why does calm feel like something is about to happen?
Calm feels like a warning because a threat-detection system does not measure danger, it measures the probability of danger — and after years of a high base rate, the setting on that estimate is high. The clinical term is hypervigilance: sustained scanning for cues of threat, maintained at a cost, continuing after the conditions that justified it have gone. It does not switch off when circumstances improve, because it never took instructions from circumstances. It took them from evidence, over years.
What that produces in a quiet house is not relief. It is the sensation of waiting. Nothing is happening, so the system reads the nothing as the pause before something, because that is what nothing used to mean. A good relationship can therefore feel less safe than a bad one: in a bad one, the vigilance was earning its keep.
There is early experimental evidence that the calibration itself shifts, and it must be reported with its size attached. In a behavioral pilot study of 24 Italian women who had experienced intimate partner violence, compared with 24 women who had not — recruited through anti-violence centers, tested online — the survivors rated highly trustworthy male faces as less trustworthy than the comparison group did, and were less accurate at identifying fear in male faces (Silvestri, Gobbo, Pugliese, Mancini & Visco-Comandini, 2025). Twenty-four women. The authors call it a pilot, it is not a finding about you, and it measures nothing about your accuracy with a real person.
What it illustrates is the shape of the thing: the adjustment is not global and not a loss of judgment. It shows up in a particular domain, about a particular class of face. Systems trained on a specific danger stay specific. If fatigue and fog run alongside this, they have their own mechanism. If the question underneath is duration, that is a separate question again.
Why does it land on him when he hasn’t done anything?
The alarm lands on a new partner because fear learning does not stay attached to the original person. The mechanism is stimulus generalization — a response conditioned to one cue transfers to cues resembling it, along whatever dimension the nervous system happened to encode. What got encoded was rarely the man. It was a tone dropping, a door closing a particular way, a message unanswered for two hours, a sentence beginning we need to talk. The new man produces one of those on a Tuesday, innocently, and the whole system fires.
You already know this is not fair, which is the part that hurts. Women describe watching themselves do it — hearing the flatness come into their own voice, checking his phone once and hating it, testing him in a way they cannot afterwards defend. That is generalization plus a system gathering information fast, and it is not evidence of a character you caught from him.
The second half of the mechanism explains why it comes back after you thought it had gone. Safety learning does not overwrite threat learning. In the fear-conditioning literature the original association is not erased during extinction; it is “left intact as new, secondary inhibitory learning about the CS-US develops” (Craske, Treanor, Conway, Zbozinek & Vervliet, 2014). Two learnings then compete, and which one wins depends on the situation — which is why fear returns when the context changes (renewal), when time passes (spontaneous recovery), and after unrelated bad events (reinstatement).
Women who say they cannot trust anyone are almost never describing anything that general. Pressed, it turns out they trust him with the children, with money, with a hospital appointment. What they do not trust is a change in his tone at nine in the evening. The system is not indiscriminate. It is precise, aimed at the exact things that used to mean something — which matters, because a precise problem is workable and a global character defect is not.
If you are in immediate danger, call 911.
National Domestic Violence Hotline — 1-800-799-7233, 24 hours · text START to 88788 · thehotline.org
988 Suicide and Crisis Lifeline — call or text 988
Deaf, DeafBlind and hard-of-hearing callers: National Deaf Domestic Violence Hotline videophone 855-812-1001, 24 hours (a partnership between the National Domestic Violence Hotline and Abused Deaf Women’s Advocacy Services).
Calling leaves the number in your call log and texting leaves the thread in your messages. Both can be deleted afterwards. If that is not something you can do safely, a friend’s phone or a payphone is the safer route.
Does this mean I have an attachment problem now?
Attachment style describes how a person manages closeness and distance. It is not a diagnosis, not a fixed feature of anybody, and not a verdict on your character — and it has become one of the most misused ideas in popular psychology, handed out as a label and used as an accusation. A strategy is learned in a place. Yours was learned where closeness carried a cost, and it is not a discovery about who you are.
The measured picture is also far less fixed than the online version. In a study combining three longitudinal samples — 628 US participants, followed across ages 13 to 72 — attachment anxiety declined on average with age, particularly in middle and later life, and avoidance decreased in a broadly linear fashion across the lifespan; being in a relationship was associated with lower anxiety and lower avoidance (Chopik, Edelstein & Grimm, 2019). That study is about ordinary development, not about abuse, and it promises you nothing. What it removes is the assumption underneath the label: these orientations are not stable traits, they move, and relationships are one of the things that move them.
Two consequences follow. A quiz result is not a finding about you. And if what you are actually describing is intrusive memory, sustained arousal and a body that reacts before you have decided anything, that is not an attachment style at all — it is a trauma presentation, assessed and treated differently. Whether it reaches that threshold is a question about whether emotional abuse can cause PTSD, and it needs a person rather than a page.
Do I have to tell him what happened?
You do not owe anybody the account, and reframing what a disclosure is for is the most useful thing this page can offer. Telling a new partner about an abusive relationship is not a confession you owe him. It is a test you are running on him — and the result is information you cannot get any other way.
The research supports that reframe, because what predicts harm is not whether a survivor discloses. It is what comes back. In a systematic review and meta-analysis of 51 studies covering 1,871 correlations, negative social reactions to disclosure — particularly reactions involving controlling the survivor, distracting her, and treating her differently afterwards — were associated with worse psychopathology, while positive reactions did not appear to be protective at all (Dworkin, Brill & Ullman, 2019). That literature is correlational and is largely about disclosure to friends, family and services rather than to new partners. The asymmetry is the useful part: the response is the variable that matters, and the response belongs to him.
Disclosure therefore works best in graded steps rather than as one long night of everything. A small, low-cost piece of the story first — one sentence, one fact, not the file — produces a real answer about how he handles it. Does he take it as information, as a problem to solve, as something that makes you interesting, or as leverage? Does he ask a question, or need reassuring? You are not obliged to hand over everything before you know that.
One further point, and it is a serious one. If your ex still has access to you — through children, court, shared people, a phone — what you say to a new partner is not only an emotional matter.
Why is “you just have to learn to trust again” bad advice?
Telling a woman to learn to trust again is bad advice because it treats trust as a decision, and trust is not a decision. Nobody has ever successfully chosen to find somebody safe. What can be built is the second learning described above — the inhibitory one that says not in this room, not with this person — by accumulating disconfirming evidence in specific situations, not by resolving to be less careful. Telling a woman to trust is telling her to skip the evidence, which is exactly what she was trained out of doing, for good reasons.
What replaces it is narrower and more usable. Instead of do I trust him, the answerable question is what has he done, in the specific situation my system is worried about? Not: is he safe. But: what happened the last four times I said no. What happened when I was wrong, or ill, or slow, or expensive, or unavailable. Those are checkable, they accumulate, and the accumulation is the only thing that has ever moved a threat estimate.
Expect the old learning to return, because the research says it will — in new contexts, after gaps, and after unrelated bad events like a bereavement or a hospital week. A return of caution after six good months is not a relapse and not proof you were fooling yourself. It is the older, stronger learning briefly winning a competition it has been in the whole time. And if what surfaces alongside the caution is anger rather than fear, that is not a different problem and not a sign this is going badly.
What this page can tell you, and what it cannot
This page can tell you why calm registers as a warning, why the alarm lands on somebody who has done nothing, why safety learning competes with threat learning instead of replacing it, what the research says about who bears the risk in a disclosure, and why attachment labels are asked to carry far more than they can. This page cannot tell you that you will trust again, and it will not, because nobody can honestly say that to a stranger.
What can be said is narrower and holds. The caution is not the damage. It is the record of what you learned, in a place where learning it was correct. Whether it updates depends partly on what happens next and partly on things nobody controls. None of that makes the system faulty. It makes it a system.
If you would like to talk to somebody, the National Domestic Violence Hotline is 1-800-799-7233, 24 hours, or text START to 88788; its services are free and available at any hour. For anything legal, WomensLaw.org is free and organized state by state.
Calm feeling unsafe is a nervous-system problem before it is anything about your character, and it is treatable — so ask about a therapist’s training, including EMDR, which is built for memories that talking about has not touched, before you book.
You do not have to have decided anything to start.
SOURCES
- Cattaneo, L.B., Bell, M.E., Goodman, L.A., & Dutton, M.A. (2007). Intimate partner violence victims’ accuracy in assessing their risk of re-abuse. Journal of Family Violence, 22(6), 429–440. doi:10.1007/s10896-007-9097-8
- Craske, M.G., Treanor, M., Conway, C.C., Zbozinek, T., & Vervliet, B. (2014). Maximizing exposure therapy: An inhibitory learning approach. Behaviour Research and Therapy, 58, 10–23. doi:10.1016/j.brat.2014.04.006
- Dworkin, E.R., Brill, C.D., & Ullman, S.E. (2019). Social reactions to disclosure of interpersonal violence and psychopathology: A systematic review and meta-analysis. Clinical Psychology Review, 72, 101750. doi:10.1016/j.cpr.2019.101750
- Chopik, W.J., Edelstein, R.S., & Grimm, K.J. (2019). Longitudinal changes in attachment orientation over a 59-year period. Journal of Personality and Social Psychology, 116(4), 598–611. doi:10.1037/pspp0000167
- Silvestri, V., Gobbo, S., Pugliese, E., Mancini, F., & Visco-Comandini, F. (2025). The perception of trustworthiness and emotional identification in women experiencing intimate partner violence: A behavioral pilot study. Brain Sciences, 15(5), 429. doi:10.3390/brainsci15050429
- Herman, J.L. (1992). Complex PTSD: A syndrome in survivors of prolonged and repeated trauma. Journal of Traumatic Stress, 5(3), 377–391. doi:10.1002/jts.2490050305
Educational content. Not therapy, legal, or financial advice. Nothing on this page is an assessment of any person or any relationship.