The Reversal

In 1997, the psychologist Jennifer Freyd, working at the University of Oregon, gave a name to something that had presumably been happening for as long as people have been able to lie to each other about harm. She called it DARVO — Deny, Attack, Reverse Victim and Offender¹ — and the name has survived because the pattern underneath it turns out to be almost mechanically consistent, across contexts that have nothing else in common.

The sequence runs in three moves. First, denial: it didn’t happen, that’s not what happened, you’re lying. The claim is simply rejected, and rejection requires no evidence — a denial costs nothing to issue or maintain. Second, an attack on the person making the claim: she’s unstable, she’s a liar, she has a grudge, she’s trying to destroy me. The target shifts from the event to the witness — because if the witness is discredited, the event no longer needs disproving; it stops mattering what she says happened. Third, the move that names the mechanism: reversal. Actually, I’m the victim here. She’s the one abusing me, with these accusations. The real harm is what this is doing to my reputation, my family, my life.

Freyd’s claim would be interesting rhetorical anatomy on its own. What makes it load-bearing here is that it has since been tested, not just observed. Harsey, Zurbriggen, and Freyd, in a 2017 study, found that among people who had confronted someone about a harm — from betrayals to serious abuse — roughly seventy-two percent reported that the person they confronted responded with some version of this exact sequence.² That is not a rare defensive reflex reserved for the unusually manipulative. It is a common one — the majority response that victims described running into when they confronted a wrongdoer, whether or not either party could have named the strategy.

A follow-up experiment, published by Harsey and Freyd in 2020, showed something more specific and more useful: 316 undergraduates read identical accounts of dating violence, differing only in whether the perpetrator’s accompanying statement used a DARVO pattern or didn’t.³ Reading the DARVO version, observers rated the perpetrator as less abusive, less responsible, and more credible — and rated the victim, describing the exact same underlying incident, as more abusive, more responsible, and less credible. Nothing about what happened had changed between conditions; only the shape of the denial had. That shape alone moved an independent observer’s judgment of who did what to whom.

---

At the scale of a single relationship, narrative control rarely begins at the moment of denial. By the time DARVO gets deployed, the ground has usually been prepared for months or years.

It tends to start with charm. An abuser typically establishes, early and convincingly, a reputation as the reasonable one — attentive, credible, well-liked by the people in the target’s life who will later be asked to weigh in on who’s telling the truth. That reputation is the foundation the later reversal will stand on. Against that foundation, gaslighting begins quietly: that didn’t happen, you’re remembering it wrong, you’re being dramatic. Each instance is small enough to dismiss on its own; cumulatively they loosen a person’s grip on their own memory of events. Isolation follows, or runs alongside it — family reframed as people who “don’t understand,” friends discouraged as bad influences, outside help positioned as unnecessary because “I’m the only one who really knows you.” And running underneath all of it is a steady accumulation of micronarratives: remember when you forgot our anniversary and made a scene; you always overreact; you’re the difficult one. None of these stories needs to be dramatic. Their power is cumulative, not individual — a slow accretion that, across enough repetitions, hardens into a settled fact in everyone’s mind, including sometimes the victim’s own: she’s the unstable one; he just puts up with her.

When actual abuse occurs, DARVO’s three moves are already primed to fire immediately: deny the severity — I barely touched you; attack — you provoked me; reverse — look what you made me do. And because the groundwork was laid first, the reversal doesn’t need to win an argument in the moment. It only needs to fit a story that’s already been told often enough to feel true.

The same three-move structure scales up cleanly to a group. A cult or a closed organization installs a single authorized version of reality — we are enlightened, our leader is wise, critics are threatened or jealous, and those who leave are traitors or unwell — and dissent inside that frame is treated not as disagreement but as evidence of the dissenter’s own spiritual or psychological deficiency. When abuse inside the group becomes impossible to deny entirely, the organization runs DARVO exactly as an individual would, just with more voices in unison: deny that it reflects the culture, attack the accuser as disgruntled or embittered, reverse the harm onto the mission the accusation is now supposedly damaging.

Institutions run the same sequence with considerably more infrastructure behind it. An institution controls its own public face — the press release, the official statement, the sanctioned history — in a way no individual accuser can match. It starts from a position of assumed credibility: professional standards, credentials, an institutional memory that reads as authority. It has silencing tools a lone perpetrator doesn’t: non-disclosure agreements, sealed records, the credible threat of a defamation suit against anyone who speaks. And its version of DARVO tends to unfold in layers rather than a single reversal — first “this was an isolated incident,” then “the person responsible was a rogue actor, not representative of us,” then “we have reformed since,” each layer adding distance between the institution and the harm it’s answering for.

The Catholic Church’s response to its own abuse scandal followed this exact sequence for decades: deny that the claims were substantiated, attack the accusers as opportunists seeking settlements, and reverse the frame entirely — cast the institution itself as the victim of a media-driven witch hunt. What eventually broke that reversal was not any single revelation but scale and documentation together: the accumulation of lawsuits, the 2004 John Jay Report’s institutional data, and years of investigative reporting that made the “isolated incident” framing impossible to sustain.⁴ USA Gymnastics ran a structurally identical sequence around Larry Nassar, denying that its screening processes had failed, attacking the credibility of the athletes coming forward, reversing the frame toward its own reputational damage, and what broke it was, again, scale: at Nassar’s January 2018 sentencing before Judge Rosemarie Aquilina, one hundred fifty-six women gave victim-impact statements across seven days, a volume of consistent testimony no institutional narrative could absorb. He was sentenced to up to a hundred seventy-five years.⁵

At state scale, the same three-part structure runs with the largest available amplifier attached. A government with an opposition figure to discredit tells its own story through its own media, terrorist, traitor, extremist, and positions itself as the injured party, defending national security against a threat rather than persecuting a critic. Courts, the psychiatric system, and state media can be brought into alignment with the official account simultaneously, so that the target’s version of events becomes not just disbelieved but functionally uncitable: the word of a “conspiracy theorist” against the apparatus of a state, with no contemporaneous documentation to settle the matter. Documentation of this kind of operation, when it exists at all, tends to surface only after declassification — years or decades too late to help the person it was used against.

---

Underneath all four scales, credibility gets reversed through the same three levers, applied with more or less force depending on how much institutional weight the perpetrator can bring to bear.

The first is authority. The person or institution running DARVO claims greater standing — credentials, position, longevity — and lets standing substitute for evidence. At state scale this advantage is close to total: an apparatus with courts and press offices behind it faces a target with none. At interpersonal scale it’s more modest but still present, usually dressed as a claim of superior insight — I understand this relationship better than you do; I’ve been through more than you have; even the therapist agrees with me.

The second is simplicity. The perpetrator’s account is almost always short: she’s unstable. The victim’s account is, almost by necessity, long: he isolated me, controlled the money, read my messages, spread a story about me at work, undermined every attempt I made to get help. Simple stories are easier to believe than complicated ones, not because they’re more likely to be true, but because they ask less of the listener. The victim’s complexity, which is simply an accurate description of how coercive control actually operates, reads to an untrained ear as an elaborate, faintly paranoid construction. The perpetrator’s simplicity reads as candor.

The third lever is fit with what the listener already believes. If an audience already carries some version of women are too emotional, dissidents are unstable, or cult members are gullible, the perpetrator’s account slots neatly into an existing category, requiring no revision of anything the listener already thinks. The victim’s account, by contrast, usually asks the listener to revise a category — to reconsider someone they liked, or a system they trusted — and that revision is cognitively expensive in a way that simply nodding along to a familiar story is not.

When all three levers align, the reversal completes: the perpetrator becomes the reliable narrator, the victim becomes the unreliable one, and from that point forward, any new evidence the victim produces gets filtered through the “she’s not credible” frame before it’s even evaluated on its merits. Confirmation bias does the rest of the work for free.

---

This is the exact point at which the psychiatric label, the subject of a later chapter, does its most devastating work. Without a label, a victim’s account reads as a coherent, if disturbing, narrative: he isolated me, controlled the money, sabotaged my job. Attach a diagnosis — borderline personality disorder is the clearest illustrative example — and the same account is instantly reframed: she’s delusional about being isolated, paranoid about his influence, seeing sabotage that isn’t there. Nothing in the underlying facts has changed; but every future statement she makes will now be read as a symptom rather than a report. The label doesn’t just discredit one accusation. It converts the accuser into an unreliable narrator of her own life, permanently, and lets the perpetrator believe — often, quite sincerely — that he isn’t gaslighting her at all. She’s simply unwell, and he is the reasonable one managing it.

This closes a trap: the more precisely a victim describes what has actually been done to her, the more elaborate and implausible her account tends to sound. Coercive control operates through small, deniable, cumulative acts — an object moved, a rumor placed at a specific moment, a job opportunity that quietly falls through — and a victim who has learned to track these small patterns, because tracking them is how she has survived, produces testimony that sounds, to someone without the frame, exactly like the thing a diagnosis would predict: a person connecting unrelated events into a conspiracy that exists mainly in her own mind. Accuracy, paradoxically, becomes the evidence used against her.

---

Truth, on its own, is frequently not enough. A victim can hold a video, a recording, a witness, a medical record — genuinely dispositive material — and still lose the argument, because evidence does not interpret itself. The same recorded behavior can be read three different ways depending on who is doing the framing: as abuse (he monitored everything I did), as protection (he cared about my safety), or as devotion (he paid attention because he loved her). The tape doesn’t decide which of the three is correct. The frame does, and the frame that wins is usually the one held by whoever has more standing to assert it.

The Stasi’s own surveillance files make this point with unusual clarity, because in that case the evidence was never in dispute — the East German state kept meticulous written records of its own campaigns against private citizens. What took years to establish was not the existence of the files but their meaning. When the Zersetzung archives were first made public, the initial response from some quarters was that this was simply routine security work, not a coordinated campaign of psychological destruction. The same documents, read through a different frame, eventually became the basis for a now widely accepted historical conclusion: this had been a deliberate program to dismantle people’s lives without ever laying a hand on them.⁶ The facts didn’t change between those two readings. The frame did, and it took sustained scholarship to move it.

---

Reversed narratives are not permanent, though the historical record suggests they can last for decades before they break, and they break for a specific, recurring set of reasons rather than through the gradual accumulation of individual persuasion.

Scale of testimony is the most reliable solvent: one accuser can be dismissed as a liar or an opportunist; a hundred and fifty-six, giving consistent statements over seven days in a Michigan courtroom, cannot. Documentation of the mechanism itself is nearly as powerful — an internal memo instructing staff on how to discredit a specific accuser proves intent rather than merely asserting harm, and intent is much harder to explain away. A shift in who holds institutional authority can break a reversal from outside: a new leadership team, a court ruling, a regulatory finding can overrule the old narrative by force of standing, without re-litigating every claim. Cultural change matters too — testimony that sounded paranoid before “coercive control” existed as a named, public concept can sound entirely credible once the frame exists to receive it; the words haven’t changed, only the listener’s capacity to recognize them. And the accuser herself can acquire institutional standing over time — through court testimony, published accounts, survivor advocacy — moving from “unreliable individual” to “credible witness” simply by acquiring a platform she previously lacked.

The pattern across every scale examined here holds without ever repeating itself in the actors. An individual abuser’s DARVO can be broken by one believing witness. An institution’s requires lawsuits, investigative journalism, and years. A state’s may require the truth to survive underground for a generation, waiting for declassification to do the work testimony alone cannot. What decides whether a lie or a true account gets to stand as history is never the move itself — it is the same move every time, run by hands that in almost every case here never met. It is what stands behind the move: for the abuser, the one person who finally believes her; for the state, a file that outlives the government that sealed it.

---

Underneath every version of the reversal sits a harder epistemic problem, and how this book answers it governs everything that follows. A reader’s instinct — that just because something isn’t visible doesn’t mean it isn’t happening — is logically correct, but the rigorous version of it cuts in the opposite direction too. Invisibility is the design goal of a well-run covert operation, which means the absence of visible evidence is genuinely uninformative: equally consistent with nothing having happened and with something having been done well. Treating “no visible evidence” as evidence of absence is a category error for a phenomenon engineered specifically to leave none. But the same invisibility cuts both ways: the absence of disproof cannot be read as proof of presence either, and this is the exact terrain where a true account and a delusional one are described in identical words — precisely the cover the reversal depends on. The resolution is neither belief nor disbelief. It is reconstruction: an invisible mechanism made visible through the traces it cannot avoid leaving — financial records, metadata, timelines, communications logs, and patterns across an entire network that no single deniable act reveals but the aggregate does. That is forensics. So the argument can never rest on “it’s invisible, therefore it’s happening” — the conspiracist’s move — or on “it’s invisible, therefore it isn’t” — the gaslighter’s move. Both are the identical error wearing opposite conclusions. The honest third path is the only one that does justice to a true victim and survives a genuine skeptic at once.

Notes

1. Jennifer Freyd, University of Oregon, coined DARVO (Deny, Attack, Reverse Victim and Offender) in 1997. https://www.jjfreyd.com/darvo↑

2. Harsey, Zurbriggen & Freyd (2017): among survey respondents who had confronted someone about a harm, roughly 72% reported that the person they confronted responded with a DARVO pattern (all three components — denial, personal attack, and reversal). Sarah J. Harsey, Eileen L. Zurbriggen, and Jennifer J. Freyd, “Perpetrator Responses to Victim Confrontation: DARVO and Victim Self-Blame,” Journal of Aggression, Maltreatment & Trauma 26, no. 6 (2017): 644–663.↑

3. Sarah J. Harsey & Jennifer J. Freyd, “Deny, Attack, and Reverse Victim and Offender (DARVO): What Is the Influence on Perceived Perpetrator and Victim Credibility?” Journal of Aggression, Maltreatment & Trauma (2020) — 316 undergraduates rated identical dating-violence accounts differently depending on whether the perpetrator’s statement used a DARVO pattern; DARVO exposure reduced the victim’s perceived credibility. The separate ~72% “all three DARVO components” figure comes from Harsey, Zurbriggen & Freyd, Journal of Aggression, Maltreatment & Trauma 26, no. 6 (2017): 644–663. https://www.jjfreyd.com/darvo↑

4. Catholic Church sexual abuse scandal and institutional DARVO response, broken by accumulated lawsuits, the 2004 John Jay Report, and investigative reporting. John Jay College of Criminal Justice, The Nature and Scope of the Problem of Sexual Abuse of Minors by Catholic Priests and Deacons in the United States (United States Conference of Catholic Bishops, 2004).↑

5. Larry Nassar / USA Gymnastics case: at Nassar’s January 2018 sentencing before Judge Rosemarie Aquilina, 156 victims gave impact statements over seven days; Nassar was sentenced to up to 175 years. https://www.cnn.com/2018/01/24/us/larry-nassar-sentencing. The institutional-response pattern described in the text — denial that screening had failed, attacks on the credibility of athletes coming forward, and reframing around the organization’s own reputational damage — is documented in the independent investigation commissioned by the United States Olympic Committee (Ropes & Gray, Report of the Independent Investigation: The Constellation of Factors Underlying Larry Nassar’s Abuse of Athletes, December 2018) and in the U.S. Senate Commerce subcommittee’s investigative report, The Courage of Survivors: A Call to Action (July 2019).↑

6. Stasi Zersetzung surveillance files: initially characterized by some as routine security work, later established through sustained scholarship as a documented program of psychological destruction against private citizens. Stasi Zersetzung as codified in MfS Directive 1/76 (1976); Jens Gieseke, The History of the Stasi: East Germany’s Secret Police, 1945–1990 (Berghahn Books, 2014).↑