This is Part 2 of two. Part 1: How the Field Left the Evidence is how the field left the evidence. This page is the correction: how memory science, courts, and practice guidelines pulled psychiatry back onto ground that can still be checked.
The climb was incomplete. Surveys still find that many therapists endorse unconscious repression even when they also agree that memory is reconstructive. Popular books still sell the body as a locker for sealed scenes. The point of this article is not that the fight ended. The point is that the field finally wrote a kill-switch. A unit that will not use that switch is choosing the 1988 funnel on purpose.
Childhood sexual abuse remains real, common, and under-reported. The correction did not require anyone to stop taking a trauma history. It required the field to stop paying for a technique whose success condition was a new memory.
What actually reversed the method
Three kinds of pressure arrived at once, which is how institutions usually move.
The lab would not certify the vault.
Elizabeth Loftus and Jacqueline Pickrell’s “The Formation of False Memories” (Psychiatric Annals, 1995) — the “lost in the mall” study — and the literature that followed did not prove that every delayed recollection is false. They proved something a clinic can use: ordinary suggestion can produce a detailed, confidently held report of an event that did not happen. A 2023 preregistered replication coded false memory in 35 percent of participants, higher than the original 25 percent. Once that is on the table, hypnosis, guided imagery, and a leading interview are no longer neutral tools. They are interventions that change the object they claim to retrieve. Richard McNally’s Remembering Trauma (Harvard, 2003) collected the laboratory and clinical case against the sealed-vault model.
The courts stopped treating the session as a discovery document.
Malpractice actions in the 1990s made the method expensive. In May 1994 a Napa County jury awarded Gary Ramona $500,000 against his daughter’s therapists and medical center — among the first times a third party, not the patient, recovered damages from recovered-memory treatment. In 1997 insurers for Rush-Presbyterian-St. Luke’s and associated psychiatrists paid Patricia Burgus $10.6 million. Elizabeth Gale’s related Illinois case settled for $7.5 million. The legal question was not “did abuse ever occur in America.” It was whether a hospital or a therapist could create a scene and then send it into a family as a finding. When juries and insurers priced that sequence, programs that had marketed excavation quietly dropped the product.
Professional bodies wrote the refusal into print.
In 1997 the Royal College of Psychiatrists told members not to use memory-recovery techniques to unearth abuse of which the patient had no memory — including drug-mediated interviews, hypnosis, regression, guided imagery, “body memories,” literal dream interpretation, and journaling. The Brandon working-party review, published in the British Journal of Psychiatry in 1998, went further: memories recovered after long amnesia, especially by extraordinary means, have a high probability of being false. The authors were careful about the sentence the field needed: this conclusion must not be used to dodge the recognition and treatment of children who were actually abused. That pairing — take abuse seriously, do not excavate a blank — is the climb.
American organizations wrote the same operational limit. The American Psychiatric Association (1993 / restated 2000, reaffirmed 2013): complete an evaluation first; keep a neutral stance; do not prejudge that abuse is or is not the cause; when a new memory has no corroboration, treat historical certainty as optional. The American Psychological Association’s working-group conclusions: most people who were sexually abused as children remember all or part of what happened; it is also possible to construct convincing pseudomemories; processes that produce accurate versus inaccurate recollection are not fully known. The AMA’s 1994 line, reported alongside those statements, was that recovered memories of childhood sexual abuse are of uncertain authenticity and should be subject to external verification.
Do not use hypnosis or drug-assisted interviews as a search engine for a childhood crime. Do not treat a symptom list as proof of a sealed event. Write what you asked.
What “climbed back” means in a protocol
The field did not climb back by discovering that trauma is unimportant. It climbed back by changing the question.
Old question: What buried scene explains this adult?
Current question: What symptoms are present, what does the patient already report, and which treatment has an outcome measure?
Prolonged exposure, cognitive processing therapy, and trauma-focused CBT do not require a recovered vault. They work with the memories and beliefs the patient already has. The target is avoidance, intrusion, and the meaning the patient assigns to what is already in reach. If the score does not move, the plan changes. That is a falsifier. Recovered-memory practice did not have one.
ICD-11’s complex PTSD and DSM-5-TR’s trauma-related chapter are also part of the climb, with a caveat. New labels can become a new vault if staff treat them as permission to hunt for an origin story the intake did not contain. A diagnosis is a workup, not an excavation license.
Reconstructive memory is the operational rule. People remember. They also compress, blend, and fill gaps. Stress changes encoding and retrieval. Delayed disclosure happens. None of that licenses a therapist to supply the missing frames and then call the result a file.
What a unit writes so the first map can die
Name the techniques that are off-protocol.
Hypnosis, sodium amobarbital interviews, and guided imagery used to produce a first memory of abuse do not belong on a psychiatric service. If a contractor wants to use them, the contract should say no.
Separate disclosure from recovery.
“I have always known and I am telling you now” is a disclosure. “I had no memory until session six, after we used a script” is a product of treatment. Those two sentences require different next steps. Only the first one can be treated as the patient’s incoming history.
Keep trauma treatment tied to a scale.
PCL-5, PHQ-9, GAD-7, sleep, substance use. If the protocol cannot say what would make it wrong at week eight, it is still a vault theory with better manners.
Do not let moral urgency replace collateral.
Asking a partner, a pediatric record, or a prior clinician for dates is not betrayal. It is how a workup stays a workup. Families who deny an allegation are not automatically right. They are also not automatically contaminating the case by existing.
Allow a second reader who does not need the origin story.
A telepsychiatry attending who was not in the first trauma group can ask whether the plan is treating symptoms or completing a narrative. Distance does not create honesty. A contract that lets that person write “the first hunt found nothing — treat what is here” does.
FasPsych’s role is that second reader when a facility wants it. Licensed psychiatrists and psychiatric nurse practitioners use the facility record, score the facility instruments, take the history the patient brings, and stay on the case when the first story is incomplete. They do not run an excavation the medical director did not order.
The climb is unfinished on purpose
Belief in repressed memory did not vanish when the guidelines landed. Otgaar, Howe, Patihis, Loftus, and colleagues (2019) found the belief still common — 58 percent in the samples they reviewed — and higher among some clinical psychologists than in the 1990s. They argue that dissociative amnesia often functions as a substitute term. Ethan Watters noted in 2022 that many trainees have no living memory of the recovered-memory years. The claim also moved into self-help language and a public appetite for a body that stores what the mind will not say. That is a reason to keep the kill-switch in the policy binder, not a reason to pretend the 1990s never ended.
A field that climbed back can still slide. The slide looks like this: a new popular book, a new technique that “unlocks,” a new staff member who treats missing recall as a positive test. The correction is not a mood. It is a list of things the unit will not buy.
Key takeaways
- The correction was lab plus court plus guideline, not a change of heart.
- Reconstructive memory is an operational rule: a new scene produced by a technique is not, by itself, a finding.
- Evidence-supported trauma therapies treat the memories and symptoms already in the room. They do not require a vault.
- Disclosure and recovery are different events. Write which one you have.
- A program that cannot name a banned technique has not climbed anywhere. It has only waited.
FAQ
Did psychiatry admit it was wrong?
Parts of it wrote guidelines that banned the methods. That is the form institutional correction usually takes. Individual clinicians still vary.
Does reconstructive memory mean patients lie?
No. It means memory is a report that can be accurate, partial, or wrong, and that interview method changes the report. Lying is a separate question.
What treatments replaced recovered-memory work?
For PTSD and related presentations: prolonged exposure, cognitive processing therapy, trauma-focused CBT, medications when indicated, sleep and substance treatment. None of those start by asking the patient to produce a first memory of a crime.
How is this different from Part 1?
Part 1: How the Field Left the Evidence is how the field left. This page is what it installed instead — and what a hospital should still refuse.
How does FasPsych fit?
Providers can complete evaluations, medication management, and follow-up on the facility’s platform, score the facility’s instruments, and decline excavation techniques that the evidence did not survive. Call (877) 218-4070 or use the partner form.
Write the refusal down
The Memory Wars ended, in practice, when courts and colleges made excavation expensive and trauma care found protocols that do not need a new past. How the method took hold is in Part 1: How the Field Left the Evidence.
A unit that still treats a blank recall as an unfinished retrieval has not learned the 1998 sentence. The sentence is simple. Take abuse seriously. Do not manufacture the file.
If the lecture about hidden origins is more important than the symptoms that can be scored this month, do not hire another consultant to reprint the lecture. Staff a clinician who is allowed to treat what is already in the room.
Partner with FasPsych when the missing piece is a licensed prescriber who will use your record, write the source of the narrative, and keep the plan tied to a scale. Partner with us or call (877) 218-4070.