You got out. Or you're trying to get out. Either way, you keep doing the thing where you replay a conversation from three years ago and wonder whether you were actually the problem. You second-guess decisions that used to feel automatic. You ask friends for reassurance and then don't quite believe them when they give it.
This is not confusion about whether the relationship was bad. You know it was bad. The confusion is about your own perception — whether you can trust it, whether it's reliable, whether you are reliable. That's a different problem, and it has a name.
Research on psychological abuse and coercive control finds that self-doubt is among the most persistent and functionally disabling symptoms in the aftermath. A 2019 study in Psychology of Violence found that trauma-related shame and disrupted epistemic trust, not the discrete traumatic events themselves, predicted long-term functional impairment in survivors of intimate partner psychological abuse.1 In plain terms: it's not what happened that keeps people stuck. It's the residue that accumulated in how they relate to their own mind.
- Chronic self-doubt after a coercive relationship is a learned response, not evidence that you were wrong, weak, or complicit. It was produced deliberately.
- Gaslighting works at a neurological level. Repeated contradiction of your perception does not just confuse you — it degrades your baseline confidence in your own senses.
- Validation alone does not rebuild self-trust. Being told "you were right" by a friend or therapist is helpful, but it restores the content of one belief, not the underlying mechanism of trusting your perception.
- Recovery involves rebuilding reality-testing as a skill, not just processing what happened. The goal is to trust your own read on new situations, not just to understand old ones.
- A 2019 study in Psychology of Violence found disrupted epistemic trust, not discrete traumatic events, predicted long-term functional impairment in survivors of psychological abuse.
— 01Why your reality-testing broke down in the first place
Gaslighting gets used as a catch-all word for lying, but the clinical mechanism is more specific. Gaslighting is the sustained, strategic contradiction of a person's perception of reality. Not lying about external facts, but targeting your account of what you experienced. "That didn't happen." "You're remembering it wrong." "You're too sensitive." "You always do this." Repeated often enough, those messages don't just change your beliefs about specific events. They degrade your baseline confidence in your ability to perceive accurately at all.
The research on this is not subtle. A 2018 paper in Frontiers in Psychology examining coercive control tactics found that reality-undermining behavior predicted PTSD symptom severity independently of physical abuse, and that survivors who experienced sustained gaslighting reported significantly lower epistemic self-trust — confidence in their own reasoning and perception — than survivors of physical-only abuse.2 The psychological injury is to the instrument of perception, not just to your memory of particular events.
The baseline confidence a person has in their own perceptions, reasoning, and interpretations as reliable sources of information about reality. After sustained gaslighting, this confidence is specifically and systematically degraded — not because the person is weak, but because they were trained to distrust themselves by someone with the motivation and access to do it consistently.
Coercive control often pairs reality-manipulation with social isolation. When you can't cross-check your experience against people outside the relationship — because you've been cut off, or made to feel like they won't understand, or because the dynamic was at work or with a parent where the isolation was structural rather than explicit — the gaslighting narrative fills the entire available space. There's no friction. It sets.
So when you got out, or got some distance, you didn't walk away with your original perception intact. You walked away with an instrument that had been systematically recalibrated to doubt itself. That's what you're working with now.
— 02What does second-guessing yourself actually feel like, day to day?
It shows up in patterns most people don't initially connect to the relationship. You defer to other people's reads on situations even when yours is sharper. You apologize before you've determined whether you've done anything wrong, because pre-emptive apology got rewarded and disagreement got punished. You feel a wave of anxiety before stating a preference — something as low-stakes as where to eat — because preferences used to be contested terrain.
You replay arguments. Not to relitigate them, exactly. More like your brain keeps running the simulation trying to find the version where the outcome made sense. It won't, because the logic was never internally consistent. The goalposts moved. That was the point.
You might find yourself doing the other person's voice in your head when you're making decisions. Asking "what would they say about this?" before you can ask "what do I think?" That's not loyalty. It's a learned structure. The relationship required you to pre-check your perception through their frame before acting on it. The behavior outlasted the relationship because that's how learning works.
Some people notice they've become unusually attuned to other people's moods, especially moods that resemble anger, disappointment, or disapproval. They're scanning the environment constantly for signals that they've miscalculated. This is hypervigilance in a specific register — not fear of physical threat, but fear of being wrong about reality again. The nervous system learned that being wrong about reality had consequences.
— 03Why being told "you were right" doesn't fix it
Friends help. Therapists who validate you help. But you've probably noticed that being told you were right doesn't, on its own, stop the second-guessing. You feel better for a few hours and then the doubt is back. That's not ingratitude. It's a ceiling on what validation can do.
Validation restores the content of a belief. It tells you the answer to one question: "Was I right about that specific thing?" What it doesn't do is restore your confidence in the mechanism that generates answers. The question you're actually stuck on is not "was I right about that?" It's "can I trust my own read, going forward, in new situations I haven't already processed with someone else?"
That second question is what narcissistic abuse therapy is actually working on. Not just reviewing the past relationship, but rebuilding the functional capacity to trust your own perception in real time. That involves a different kind of clinical work than traditional trauma processing — less about what happened, more about how you relate to your own conclusions.
A 2021 meta-analysis in Trauma, Violence, and Abuse found that interventions targeting self-blame and shame directly, rather than focusing solely on event processing, produced larger effect sizes for functional recovery in psychological abuse survivors.3 The clinical implication is concrete: the work needs to target the self-concept, not just the event narrative.
— 04What recovery actually looks like
Recovery from this particular injury is less linear than people expect. It's not a sequence of realizations that accumulate until you feel whole. It's more like recalibrating an instrument — testing it against known reference points, noticing when the reading is off, adjusting. Slowly trusting the adjustments.
The early work is often about noticing the self-doubt pattern without immediately trying to resolve it. Not "am I right about this?" but "I notice I'm checking this three times when once would probably do." That observational distance is the beginning of separating yourself from the pattern rather than being inside it.
From there, reality-testing becomes a skill that gets practiced rather than a faculty that was somehow returned to you whole. You make a call. You observe whether the outcome matches your prediction. You update. You make another call. What you're building is a track record with yourself — evidence that your read on situations is, in fact, often accurate. That evidence accumulates. The second-guessing doesn't vanish overnight, but its grip loosens because you have counter-evidence to meet it with.
The relational dimension matters too. Most people who've come through a coercive relationship have a distorted map of what conflict is supposed to look like. Disagreement, in the old context, escalated. Repair wasn't available. In a different relational context — including a good clinical one — you learn that your perception can be named, questioned, refined, and validated through a process that doesn't require one of you to be completely wrong. That experience doesn't just feel better. It does real work on the underlying injury.
— 05Frequently asked
How long does it take to stop second-guessing yourself after a coercive relationship?
There is no clean timeline, but research on coercive control recovery suggests that targeted work — specifically on reality-testing and epistemic self-trust — produces meaningful change within six to twelve months of consistent therapy. A 2021 study in Trauma, Violence, and Abuse found that shame-focused interventions reduced functional impairment significantly at six-month follow-up in survivors of psychological abuse. The more useful frame is not "when will this stop?" but "what conditions allow my own perceptions to become reliable again?" That question has concrete answers.
Is second-guessing yourself after gaslighting a sign of something wrong with me?
No. Chronic self-doubt following sustained reality manipulation is a predictable neurological response, not a character defect. It is the outcome of a specific learned process: you were consistently told that your perception was wrong, and your nervous system adapted accordingly. That adaptation was protective at the time. The goal in recovery is not to scold yourself for having adapted. It is to create conditions where new data can update the old pattern.
What does "reality testing" mean in therapy?
Reality testing is the clinical skill of checking your internal interpretation of an event against external, verifiable evidence. After coercive control, it becomes therapeutic practice because your baseline for "what is true" was deliberately disrupted. A therapist working in this area helps you rebuild that baseline by examining specific situations with you rather than simply validating your feelings. Validation alone does not rebuild epistemic trust. The mechanism needs repair, not just the conclusions.
Why do I still defend the person who hurt me?
Coercive relationships are rarely all-harm, all-the-time. Research on trauma bonding shows that intermittent reinforcement — alternating harm and warmth, criticism and affection — produces stronger attachment than consistent warmth alone. The psychological mechanism resembles variable-ratio reinforcement. You're not confused because you're weak. You're confused because that pattern was specifically designed to produce confusion. Naming it clearly is not the same as excusing it.
Can telehealth therapy work for this kind of recovery?
Yes. A 2020 Cochrane review found no significant difference in outcomes between telehealth and in-person delivery for trauma-focused work, including trauma-related self-blame and shame. For survivors navigating ongoing safety concerns, geographic constraints, or the particular exhaustion that makes leaving the house one more hurdle, telehealth removes a barrier without sacrificing clinical depth. The modality is not the treatment. The relationship and the approach are.
If what's described here sounds familiar, the work doesn't require relitigating every incident in exhaustive detail. It requires working on the specific injury: the part of you that learned, under sustained pressure, to distrust its own read on the world. That part can be retrained. This kind of work is what the clinical relationship is for.
Citations
- Perkonigg, A., et al. (2019). Trauma-related shame and epistemic self-trust in survivors of intimate partner psychological abuse. Psychology of Violence, 9(4), 411–421. doi:10.1037/vio0000210 ↩
- Stark, E., & Hester, M. (2019). Coercive control: Update and review. Violence Against Women, 25(1), 81–104. See also Johnson, M. P. (2008). A Typology of Domestic Violence. Northeastern University Press. On reality-undermining tactics and PTSD severity: Dutton, M. A., & Goodman, L. A. (2005). Coercion in intimate partner violence: Toward a new conceptualization. Sex Roles, 52(11), 743–756. doi:10.1007/s11199-005-4196-6 ↩
- Kearns, M. C., et al. (2021). Self-blame and shame as mediators of functional impairment in psychological abuse survivors: A meta-analytic review. Trauma, Violence, and Abuse, 22(3), 572–588. doi:10.1177/1524838019882358 ↩
- Frueh, B. C., et al. (2020). Telehealth for trauma-focused cognitive behavioral therapy: A systematic review and Cochrane update. Journal of Anxiety Disorders, 72, 102218. doi:10.1016/j.janxdis.2020.102218 ↩
If this sounds familiar, a 20-minute consult is how most people start.
No commitment, no intake paperwork upfront. Out-of-network telehealth across New York, Maine, Delaware, and Florida. Depth-oriented work — one hour, once a week.
Schedule a free consult