It sounds like fodder for an ABC Sunday night made-for-tv-movie: A young man or woman is suffering from depression and seeks therapy. The concerned therapist, hyper-enlightened about the high incidence of sexual abuse against children, immediately (almost telepathically) suspects that the patient's depression is a symptom of post-traumatic stress disorder arising from repressed memories of childhood sexual abuse. The therapist learns that the patient has recurring nightmares involving a shadowy, unidentifiable, vaguely threatening figure. The therapist believes these nightmares are the key to the patient's repressed memories, and assists the patient in unlocking the images of the dream. After weeks of relaxation techniques and gentle guidance from the therapist, the patient "remembers" that the images in the dream are images of sexual abuse. After this breakthrough, the patient then further "remembers" that the dreamed-of sexual abuse actually happened to him or her as a child twenty years ago, and that the perpetrator was your client. The patient tearfully thanks the therapist and proceeds to the district attorney's office, from which a criminal complaint against your client is promptly filed.

In the past several years, a number of both criminal and civil cases alleging sexual assault or abuse have been brought based on so-called "recovered memories" of abuse that allegedly occurred years (sometimes decades) earlier. Often these distant memories are the only potential evidence of the charged abuse. In order to substantiate these abuse claims at trial, the criminal prosecutor or civil plaintiff must present evidence of recovered memory theory, a psychological proposition that people repress memories of traumatic events, and, more importantly, that people can be assisted in recovering those memories accurately years after the events allegedly occurred.

This paper will provide defense lawyers with the beginnings of an argument for preventing the admission at trial of both (1) expert testimony about recovered memory theory, and (2) the content of allegedly recovered memories (derived from recovered memory therapy). This argument is designed to supplement an evidentiary hearing at which experts in psychology are presented to testify about the defects of recovered memory theory, and the clinician who extracted the alleged memories at issue is cross-examined about his or her methodology. Because the theory that suppressed memories recovered through therapy are factually accurate is an unreliable theory generally, and because this theory is often unreliable as applied to elicit memories from a particular patient, trial courts should exclude this evidence under the standards set forth in both Frye v. United States and Daubert v. Merrell Dow Pharmaceuticals, Inc.²

For nearly three quarters of a century, the test for admitting scientific evidence in the majority of jurisdictions was "the Frye test" articulated by the D.C. Circuit Court in Frye v. United States.³ The Frye court held that testimony deduced from a scientific theory will be admitted at trial only if the theory is "sufficiently established to have gained general acceptance in the particular field in which it belongs."⁴

Seventy years after the decision in Frye, the United States Supreme Court denounced its use in federal courts, holding that Congress intended for the Federal Rules of Evidence to supersede or "trump" Frye's dictate and impose a more flexible admissibility test, in Daubert v. Merrell Dow Pharmaceuticals, Inc.5 Consistent with Frye, the Court held that under the Federal Rules, federal courts must still fulfill a gatekeeping role, and "must ensure that any and all scientific testimony or evidence admitted is not only relevant, but reliable."⁶ However, Frye's "general acceptance" test was deemed to be only one indicia of reliability, albeit "an important factor."⁷ In addition to that factor, the Court suggested that judges ruling on admissibility consider (1) whether the scientific theory or technique at issue "can be (and has been) tested"; (2) whether the theory or technique "has been subjected to peer review and publication"; and (3) what the "known or potential rate of error" is for the theory or technique.⁸

Although Daubert was based on an interpretation of the Federal Rules of Evidence, some state jurisdictions have since renounced their common law adoption of the Frye test in favor of Daubert's more flexible admissibility guidelines, particularly where the state's evidentiary rules were modeled after the Federal Rules.⁹ The Daubert opinion provides more detailed and useful guidance for evaluating the reliability (and thus the admissibility) of expert evidence than does the Frye opinion, and might be urged on non-Daubert courts for this reason. On the other hand, practitioners in states that have retained the Frye test can make use of at least the general acceptance analyses contained under that prong in Daubert-based opinions.

Just this term, the Supreme Court clarified that "Daubert's gatekeeping obligation . . . applies not only to "scientific" testimony, but to all expert testimony," in Kumho Tire v. Carmichael.¹⁰ This decision holds particular significance in recovered memory cases, where the proffered expert testimony is psychological, rather than "scientific" in the traditional sense. Prior to the decision in Kumho Tire, criminal prosecutors and civil plaintiffs could argue that the reliability inquiry set forth in Daubert did not apply to anything but the "hard sciences," and thus recovered memory syndrome was immune from a rigorous Daubert attack.¹¹ This argument should not meet with much success in the wake of Kumho Tire.

In setting forth the above-cited list of factors for determining the admissibility of scientific evidence, the Daubert Court cautioned that "many factors will bear on the inquiry, and we do not presume to set out a definitive checklist or test."12 The factors enumerated in Daubert all concern the general reliability of the theory at issue, without regard to the theory's application to the specific case at hand. A strong argument can be made that in cases involving psychological evidence offered to prove the occurrence of a past act, two additional factors should be considered concerning the reliability of the theory as applied in the particular case before the court.¹³ Those factors are (1) whether the psychological technique used to elicit the statements (in this case, the memories) that the past act occurred was overly suggestive; and (2) the existence of corroborating evidence. These considerations have been deemed valuable in other situations where psychological influences may have rendered testimony about the occurrence of a past act unreliable, such as in cases involving possibly-tainted eyewitness testimony, and cases involving the testimony of children making accusations of sex crimes after overly-suggestive interviews.¹⁴

In sum, a modified Daubert approach to whether proffered recovered memory evidence is admissible would take into account the following factors: (1) whether recovered memory theory can be (and has been) tested; (2) whether recovered memory theory has been subjected to peer review and publication; (3) what the "known or potential rate of error" of recovered memories is; (4) whether recovered memory theory has achieved general acceptance in the field of psychology; (5) whether the psychological technique used to elicit the alleged victim's recovered memories was unduly suggestive; and (6) the existence of other evidence to corroborate the alleged victim's claims of abuse.¹⁵

Reported cases considering the reliability of recovered memory syndrome have arisen primarily in two contexts. The majority of cases are concerned not with the admissibility of such evidence at trial, but rather with whether the statute of limitations for civil suits arising from abuse allegations was tolled during the period the plaintiff's memory was allegedly suppressed.¹⁶ While these cases offer a helpful analysis of the recovered memory debate, they are not directly relevant to the question of whether recovered memory should be admitted at trial ? a question raising reliability issues not necessarily pertinent to a statute of limitations decision.

The few reported cases considering the admissibility of recovered memory theory at trial offer inconsistent results. For instance, the Supreme Court of New Hampshire recently reviewed the available literature on recovered memories and found that "the phenomenon of recovery of repressed memories has not yet reached the point where we may perceive these particular recovered memories as reliable"; thus evidence of recovered memories was deemed inadmissible under a modified Daubert test in that jurisdiction.¹⁷ In contrast, the Federal District Court for the Eastern District of Michigan somewhat less recently applied the Daubert test to hold that one proposed expert demonstrated sufficient support for recovered memory theory in the field of psychiatry to allow her to present that theory to a jury.¹⁸

In light of these conflicting results, defense lawyers can urge courts in undecided jurisdictions to examine the issue anew and make independent determinations of the admissibility of recovered memory evidence in particular cases. In making this determination, courts should place the burden of proving reliability on the proponent of the proffered evidence. Although they reached different results, both the New Hampshire and Michigan Federal District Court opinions mentioned above were consistent in requiring the proponent of recovered memory evidence to bear the burden of proving its reliability.¹⁹

Recovered memory theory fails to pass the general reliability test when all four of the original Daubert factors are weighed:

One of the most oft-cited critiques of recovered memory syndrome is that its validity cannot be tested. As one author notes, "[i]t has not been possible to demonstrate any clear link between clinical accounts of trauma, and the neurobiology of memory," and, moreover, "there is no reliable means of distinguishing a true memory from an illusory one other than by external confirmation."²⁰ As will be demonstrated more thoroughly below, memories can be distorted from suggestion, and according to one expert, "once activated, the manufactured memories are indistinguishable from factual memories."21 Psychiatry's inability to test the accuracy of recovered memories weighs against admitting such memories and their related theories at trial.

Proponents of recovered memory theory argue that recovered memory theory has been "tested" for accuracy, with favorable results. When these studies are cited, it is essential to recognize that most of them concern the question of whether memories can be or are ever repressed. The question of whether such repressed memories can be accurately retrieved through therapy has not been addressed by these studies to any satisfactory extent. The available studies of repressed memories, and the critiques of these studies, include the following:

The Briere & Conte study.

According to this study, 59% of 450 adult clinical clients who reported having been abused had repressed memories of that abuse at some time.²² This conclusion was based on the subjects' answers to the following question:

During the period of time between when the first forced sexual experience happened and your 18th birthday was there ever a time when you could not remember the forced sexual experience?²³

Of course, this question -- "was there ever a time when you could not remember" -- might have been interpreted differently by different subjects. For example, some might have interpreted it to mean "was there ever a time when you could not bring yourself to remember the abuse?" This study has been criticized on precisely this ground.²⁴ Additionally, since all of the subjects of this study were in therapy, questions remain as to whether the treating therapists had, whether consciously or not, been suggestive in their belief that repression is common, or, indeed, in their belief that these patients had actually been abused.²⁵

Most noteworthy, even if Briere and Conte are correct that 59% of abused people repress memories of the abuse, this study does not tell us anything about whether those memories can be accurately retrieved through therapy, or whether "memories" retrieved through therapy can be tested for accuracy.

The Loftus study.

In the Loftus study, a much lower estimate of the rate of repression was demonstrated, when 18% of 100 outpatients for substance abuse claimed to have once forgotten about childhood sexual abuse that they later remembered.²⁶ But again, this study only offers a percentage of people who might repress memories of abuse, and does not tell us anything about whether repressed memories can be accurately retrieved through therapy, or whether retrieved memories can be tested for accuracy absent external confirmation.

The Williams study.

At first blush, this study appears to be a convincing test of accuracy because it is based on a sampling of 200 women who were actually treated for sexual assault in hospital emergency rooms at a young age, thereby eliminating the question of whether the subjects were actually abused.²⁷ These women were asked 17 years later whether they remembered the abuse for which they had been treated, and the results showed that "38 % did not recall the abuse or chose not to report it."²⁸ These results have been criticized on grounds that the study was not conducive to eliciting honest responses, and that some of the subjects had been so young (as young as 10 months of age) when they were treated that their failure to recall was more likely ordinary forgetting than clinical repression. Most important for the criminal defense lawyer seeking to exclude recovered memory evidence at trial, the Williams study, again, does not address the question of whether repressed memories can be accurately retrieved through therapy.

The Herman/Shatzow study.

The only study this author located that attempted to test the validity of memories of abuse was the Herman/Shatzow study, involving only 53 subjects.²⁹ These subjects were all in therapy as "incest-survivors"; only some of them had demonstrated a previous memory loss, and only about 14 or 15 had "severe memory defects."³⁰ When the subjects were asked to corroborate their memories, 14 of them were unable to find corroborating evidence of the abuse they remembered. The main flaw of this study is that it does not show whether the 14 who could not corroborate the abuse were the same 14 who had suffered from severe memory defects.³¹ Additionally, this study does not indicate what methods were used to help these women recover their memories. Consequently, it cannot be relied on as an adequate test of whether or how people who have completely repressed memories might accurately retrieve those memories.

The proponent of recovered memory evidence may be expected to offer these or similar studies of repressed memories to the court. However, the critiques of these studies are widespread and soundly argued. More importantly, these studies fail to test the reliability of the theory that the proponent wishes to present: that repressed memories can be accurately retrieved. The evidence under this factor weighs in favor of exclusion, the above "tests" notwithstanding.

Journals of psychology and psychiatry are replete with articles both advancing and critiquing recovered memory syndrome. Although this theory has been subjected to peer review and publication, a great deal of that peer review and publication has consisted of harsh criticism of the theory, and this factor is not alone enough to justify admitting evidence derived from this much-criticized theory at trial, as all other factors weigh in favor of exclusion.

Although, as argued above, the validity of any particular memory is not capable of being tested (absent external corroboration), many studies have shown that "memory is malleable even for life's most traumatic experiences," and memories of particular events can be invented whole cloth with a modicum of suggestion or influence.³². Elizabeth Loftus, a psychologist nationally recognized for her expertise in the area of suggestibility and memories, has noted that "there are hundreds of studies to support a high degree of memory distortion."³³ Those studies include the following three experiments:

The shopping mall experiment.

In this experiment, adult subjects were queried by family members about whether they remembered getting lost in a shopping mall at the age of five. Approximately 20 to 25% of those questioned remembered the incident in detail, even though their family members agreed the incident never occurred.³⁴

The ear infection experiment

In this study, college students were asked to recall childhood events which parents confirmed never happened, such as an overnight hospital stay for an ear infection. Approximately 25% of the students erroneously claimed to remember the events.³⁵

The Challenger space shuttle experiment

On the day after the Challenger space shuttle exploded, professors at Emory University asked their college freshmen students to write a description of where they were and what they were doing when the Challenger exploded.³⁶ Three years later, the professors asked the same students to recall where they were and what they were doing when the Challenger exploded; the professors then compared the statements to those made the day after the explosion.³⁷ The experimenters reached two conclusions: first, there was a high level of inaccuracy in the recollections three years later, and second, high confidence levels accompanied completely wrong recollections.³⁸

Proponents of recovered memory theory argue that experiments such as these are irrelevant because they involve the implantation or recovery of a non-traumatic memory. For obvious ethical reasons, psychologists cannot attempt to discover the potential rate of error in recovered memory therapy by attempting to convince the subjects of their experiments that they were once victims of truly traumatic events. However, some evidence that elicited memories have a high error rate has been gleaned from studies of memories in the aftermath of actual traumatic events. For instance, Loftus cites one study of children's recollections of a sniper attack on an elementary school playground.³⁹ Children who were not present during the shooting, but had been exposed to stories about it, erroneously claimed to remember witnessing the attack, in detail.⁴⁰

Although recovered memory theory is not susceptible to testing such that a "known rate of error" could be determined, its potential rate of error has been demonstrated by experiments such as these herein described. This factor weighs against admitting evidence derived from recovered memory theory at trial.

A review of the current literature reveals that recovered memory theory has not achieved general acceptance in any relevant field. The absence of general acceptance can be demonstrated by the critiques of various experts as well as by the cautionary positions of a number of professional organizations. Commentators have emphasized "the magnitude of the rift" between those who believe recovered memory syndrome should generally be considered valid, and those who believe that recovered memories are generally suspect. Others conclude that "most experimental and cognitive psychologists strongly reject the validity of recovered memories."

Statements by professional organizations include the following:

Royal College of Psychiatrists

In response to the growing debate in the United States over recovered memory theory, the Royal College of Psychiatrists formed a committed to study the theory, and issued one of the most recent reports on the debate.⁴³ In its committee report, the College condemned current practices of recovered memory therapy, announcing:

When memories are 'recovered' after long periods of amnesia, particularly when extraordinary means were used to secure the recovery of memory, there is a high probability that the memories are false, i.e., of incidents that had not occurred.⁴⁴

The College further reported that "there is . . . a striking absence in the literature of well corroborated cases of such repressed memories [of traumatic events] recovered through psychotherapy," and "individual autobiographical memory is unreliable."⁴⁵

American Medical Association

This long established professional association has concluded that "recovered memories of childhood sexual abuse [are] of uncertain authenticity . . . [and] should be subject to external verification. The use of recovered memories is fraught with problems of potential misapplication."_⁴⁶

American Psychological Association

This organization concluded in a 1994 report that "it is possible to construct convincing pseudo-memories for events that never occurred." ⁴⁷ However, the 6-member working group responsible for the report was unable to agree about the weight to be attached to this possibility.⁴⁸ According to later reports, it was the schism over this issue which led the APA to split into two organizations, the American Psychological Association and the American Psychological Society.⁴⁹

American Psychiatric Association

The American Psychiatric Association has warned that repeated questioning may provoke the reporting of memories of events that never occurred, but "[s]cientific knowledge is not yet precise enough to predict how a certain experience or factor will influence a memory in a given person."⁵⁰ It should also be noted that recovered memory syndrome is not listed as a diagnostic category in the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders IV.

Australian Psychological Association

This organization has condemned the use of recovered memories, noting that empirical research on memory does not support the concept.⁵¹

Canadian Psychiatric Association

This organization has issued a position statement calling for great caution in the acceptance of reports of recovered memory of sexual abuse without solid corroboration.⁵²

Clinicians might argue that as a group, practicing clinicians accept the theory of recovered memories. However, clinicians' use of recovered memory theory for the purpose of therapy does not render it reliable for the purpose of truth-finding in a courtroom. Critics of recovered memory theory emphasize that "some clinicians believe, as did Freud, that historical truth is not important to therapy." ⁵³ If the purpose of a clinician is merely to assist a client in coping with disturbing dreams and shadows of memories, then the role of recovered memory syndrome in therapy is wholly different from its role in a truth-seeking judicial process, and the clinician's acceptance of recovered memory theory for this limited clinical purpose is irrelevant.

This distinction is a crucial one to emphasize when arguing for the exclusion of recovered memory evidence at trial.

The United States Supreme Court recently adopted this argument in a different context, noting the difference between the use of a psychological device in the field as opposed to its use in a courtroom, in the Court's recent analysis of the admissibility of lie detector tests in court martial proceedings. In United States v. Scheffer, the Court upheld a rule banning lie detector results from evidence in court martial proceedings.⁵⁴ The defendant there had wanted to present favorable lie detector results in his defense.⁵⁵ The Court noted that the government has a legitimate interest in limiting the admission of evidence to reliable evidence, and concluded that "there is simply no consensus that polygraph evidence is reliable. To this day, the scientific community remains extremely polarized about the reliability of polygraph techniques."⁵⁶ The Court rejected the defendant's argument that the lie detector must be reliable because it is used by the government in the field, explaining:

Respondent argues that because the Government--and in particular the Department of Defense--routinely uses polygraph testing, the Government must consider polygraphs reliable. Governmental use of polygraph tests, however, is primarily in the field of personnel screening, and to a lesser extent as a tool in criminal and intelligence investigations, but not as evidence at trials. Such limited, out of court uses of polygraph techniques obviously differ in character from, and carry less severe consequences than, the use of polygraphs as evidence in a criminal trial. They do not establish the reliability of polygraphs as trial evidence.⁵⁷

Likewise, the use of recovered memory theory in counseling sessions serves a wholly different purpose than the use of this theory at trial. Patients seek counseling to cope with depression and other psychological ailments. In the courtroom, they want to invoke the law to cast judgment upon and, in criminal cases, to revoke the freedom of the accused. The consequences of judicial proceedings are far too severe to rely on the clinician's use of this theory in counseling as a measure of its reliability as trial evidence.

This distinction did not gone unnoticed by one court in its review of recovered memory evidence. The Supreme Court of Rhode Island has cautioned that, "[w]hile psychoanalysis is certainly of great assistance in treating an individual's emotional problems, the trier of fact in legal proceedings cannot assume that it will produce an accurate account of events in the individual's past."⁵⁸

The evidence in the scientific literature set forth above demonstrates that there is no general acceptance of recovered memory theory as a truth-seeking device within any relevant field, and this final Daubert factor weighs against admitting evidence derived from recovered memory theory at trial.

Finally, even if a court determines that recovered memory theory is generally reliable enough for admission into evidence, it can be argued that its application in a particular case is unreliable, as evidenced by consideration of the following two factors:

Many therapists begin with an assumption, or a "working hypothesis" that a patient has been abused even before the patient remembers any abuse. Studies have shown that a treating therapists' preconceived notions about his or her patient's past can influence the patient's memories, and "even subtle pressure ? signaling expectations or leading questions ? can easily lead to conformity."⁵⁹ The treating clinician and all available treatment notes should be thoroughly examined for this bias, as another basis for excluding recovered memories in a particular case.

Additionally, experts in some cases have claimed that they aided patients in recovering memories by asking them to describe and analyze recurring nightmares. However, this method of memory recovery has been criticized on grounds that "there is no evidence that dreams are a 'royal road' to historical accuracy, and interpretations usually reflect the training and personal convictions of the therapist."⁶⁰ Consequently, literal dream interpretation used as evidence of fact has been said to be "of dubious provenance."⁶¹ Other, less frequently relied on, methods of memory recovery that can be exposed as questionable include the use of drugs, hypnosis, age regression, and survivors' groups.⁶²

Experts will attempt to substantiate a patient's claimed memories by testifying that the memories are consistent with the patient's symptoms of depression and post-traumatic stress disorder. However, other experts caution that "sexual abuse cannot be diagnosed through use of a symptom check list."⁶³ Indeed, even the editors of the Diagnostic and Statistical Manual warn that "nonclinical decision makers should also be cautioned that a diagnosis does not carry any necessary implications regarding the causes of the individual's mental disorder or its associated impairments."⁶⁴ Therapists often fail even to consider the possibility that an event other than sexual abuse caused the patient's symptoms. But it is clear that "no one symptom or set of symptoms (either initially or long-term) is pathogenic of childhood sexual abuse," and consequently:

the practitioner should be open to the possibility that other childhood events and trauma (e.g., parental separation and divorce; family violence; significant deaths ? including suicides ? and illnesses; medical conditions requiring invasive techniques, pain, and physical immobility; serious accidents; and natural disasters) might account for the patient's posttraumatic symptoms.⁶⁵

Finally, experts will predictably testify that a patient's memories must be accurate because of the patient's own level of certainty about their accuracy. A patient's apparent level of certainty does not mitigate against the suggestive nature of recovered memory therapy. One expert points out that "[a]n enormous body of results from cognitive psychology and cognitive neuroscience demonstrates that memories are vulnerable to distortion, and that confidence and vividness do not always go hand in hand."⁶⁶ And another cautions that "[c]onfidence in one's memory does not correlate with the accuracy of the memory, nor does the detail involved in the memory or its emotional strength give any indication of its accuracy."⁶⁷

These are only some of the arguments which might be raised in a particular case to argue that therapy resulting in recovered memories of abuse, particularly when it is consistent with a treating therapist's preconceived ideas about the patient's past, is too suggestive and unreliable to support the admission of recovered memory theory into evidence.

This factor is of course completely dependent on the facts of a particular case. It should be remembered, however, that even if there is evidence to corroborate the patient's claim of past abuse, any recovered memories identifying your client as the abuser should not be admitted absent some corroboration of the identity of the perpetrator appearing in the patient's alleged memories.

Although this paper has been primarily focused on the proposed modified Daubert test for the admissibility of expert testimony, the unreliability of recovered memory theory also renders the alleged victims' testimony about recovered memories inadmissible. As the Supreme Court of New Hampshire pointed out, "[a] recovered memory that previously had been completely absent from a witness's conscious recollection . . . cannot be separated from the process, if any, that facilitated the recovery."⁶⁸ As set forth above, therapy which results in recovered memories is often unduly suggestive, and "once activated, the manufactured memories are indistinguishable from factual memories."⁶⁹ A patient's "recovered memories" are just as unreliable as the theory of recovered memory therapy from which such memories are derived. Therefore, neither the patient nor the expert should be allowed to testify about the recovered memories at issue.

Finally, if a court deems proffered recovered memory evidence admissible, defense counsel should request an order directing the expert not to testify as to whether he or she believes that the recovered memories are true. As one court has explained: "Such testimony would invade the province of the jury by vouching for the credibility of [the plaintiff] and would, in any event, be unhelpful to the jury since everything [the expert] knows about the alleged events is hearsay from [the plaintiff]."⁷⁰

The available scholarship concerning recovered memory theory strongly weighs against its admission at trial. Under either the modified Daubert approach suggested here, or the simpler Frye test, such evidence should be excluded as unaccepted by any relevant community of experts, and as unreliable for use as a truth seeking device in a court of law.