Searching For Memory: The Brain, The Mind, And The Past

Citation preview

searching ��

for ��

memory

Searching for Melllory

The Brain, the Mind, and the Past

DANIEL L. SCHACTER ... BasicBooks A DIVIJIQO'I of ,-b.rpcrCo![;nsPlob.'is«n

For III)' /l/OI1ier, H(/f(;et

Copynght 0 1996by D�lIiel L. Sch�cter. Published by l3aslcBooks. A DI"isian ofH;IrperCollms Pubhshen.lnc. All nghts reserved. PTinted III Ihe Unl1C"d SUtCS ofArnenca. No p�rt of Ihls book m�y be reproduced 111 �ny manner wh�t:socver without WTII­

It'n pl'rmi'�lon CXCCpl in tit...

c.u...

ofbncf qu ot:lf10n s embodiedin cri1-

ic�llTl1c1cs and rcvicw>. For i nfornmion. addr\'SS Ua;LcUooks. 10 East 53rd StrcCI. N...wYork. NY 10022-5299. FIRSr U)lTION

Dcslgned by Ellio(\ Heard ISlIN 0-465-02502-1 9697989900 :Iin. A� he paints, he con­ fi-onts the discrepancy bc[\'.'ccl1 the view of memory as a static repro­ duction and what his own experience is telling him. I-Ie writes:

if Illy memory ought to be an accur:ue replica of the original expe­ rience, if that was so, my p ..inting was hopcll.!ssly maccurate. It was

In this book I identify and discuss different enlble us to hold information for brief periods o( time, to learn skills and acquire habits, to recognize evcryday objects, to retain conceptu31 information, and to recollect specific evcms. Acting in concert, thest! memory systcms allow us [Q accomplish the tasks of our day-to-day lives while also supplying our intellect and emotions with ideas and feelings from the past that allow us to act with purpose and live rich emotional lives. But memory involves morc than JUSt our rcmelll­ br:mcc of things past. As we h:we come to learn that memory is not one single thing, we've opened up a whole new world of implicit. nonconscious memory that underlies our abilities to carry out efo f rt­ lessly such tasks as riding a bicycle or playing a piano, without having to direct each movemem consciously every time we attempt the t:l�k. Many of us chink of this type of memory as being stOred in our fin­ gers, but new research is uncovering that specific brain systems are involved in the non conscious effectS We now know enough about how memories are stored and retrieved [Q demolish another long-standing myth: that memories are passive or literal recordings of realiry. Many of us still see our memo­ . ries as a series of f 1mily pictures srored in the photo album of our minds. Yet it is now clear that we do not store Judglnent-free snapshots

a bad painting of a fuzzy memory. But [ preferred to think th:lt

of our past experiences but rather hold on to the meaning, sense, and

memory is never frozen, nor should it be. My painting was a suc­

emotions these experiences provided us. Although serious errors and

cessful rendering of the dynamiC memory that had simply begun

distortions occur relatively inCrequently, they furnish significant clues

with the orlgmal ewnt.., . My painting. I figurl.!d. was so vcry

abollt how we remember the past because they arise from, and pro­

accurate III its depiction of this memory th:lt it would inevitably

vide a window on, somc of the fundamental properties of our mem­

look wrong when compared to the original model.' Philosophers and writers have sought to penetrate memory's I11Ys­ teries for centuries. and scienti�{S have struggled with remembering

ory systems. One especially important such property is that we cannot separate our memories of the ongoing events of our lives from what has hap­ pened to tiS previously. Imagine that for a set time period, two people

6

Sl'llfr/dtlg for Melllory

Intro d u c t i o n

7

were tied together so thal each could witness only whar the other saw,

tOols to bear on undersranding baffiing cases of amnesia. Clinicians

rcad on1y what the other read, learn only what the other learned, and

interested in memory loss drew increasingly on the techniques and

have only the emotional experiences the other experienced. Unless

theories developed by cognitive psychologists, and used new methods

these tWO people WCfC identical personalities with identical pam, their rncl\\ories of the tiH1C period could be vastly different. Wilat has hap­

(or visualizing (he brain, Stich as magnetic resonance imaging (MRI), to provide precise characterizations of br. lin danuge in tht'ir patients.

pened to

At the same time, neuroscience made stunning progress, facilitated by

LIS

in the past dctt.!rrnines what we take OUt of our daily

encounters in life; memories arc records of how we have experienced

technical breakthroughs that allowed increasingly refined explorations

events, not replicas of the evellts themselves. Expcricnces 3rc encoded

of the brain and by the development of powcrful new theories using

by brain networks whose connections have already been shaped by

neural networks. More and more neuroscientists bCbran to relate their

previous encounters with the world. This preexisting knowledge powerfully influences how we encode and store new memories, thus

thc past few years, ncw functional neuroimaging techniques, stich as

fmdings with rats and monkeys to human memory. And during just

contributing to the nature, tex(Ure, and quality of wha( we will recall

positron emission tomography (PET scanning), have allowcd us to see

of the m01l1!.:llt.

the brain in action while people remember. Cognitive psychologists,

Not surpri�ingly, these insights and others have taught us much

clinicians, and neuroscientists are all now contributing to pathbreak­

about the vulnerability of memory-why Ollr recollections are some­

ing neuroimaging research that is providing a novel window on mem­

times pr!.:disposed to corruption by suggestive influences, and how we

ory and brain. A synthesis has emerged during the past tWO decades

sometimes distort the past for no illmlcdiately apparent reaSOH. And

that is exciting and vast in scope.

we are beginning [0 unde�tand why some memories have the power

I decided to write this book because I believe it is time to tell the

to induce us to cry, to laugh. or [0 tremble. We are still far from being

talc from the perspective of someone who has been part of it. For

able to say that we have a complete picture of how human memory

much of my career I have attempted to link cognitive psychology,

works. but after centuries with little Sllccess, we are starfing to fmd

clinical observations, and neuroscience into a cohesive approach to

places for many pieces of the puzzle.

understanding memory. I-I ere I try to paint the big picture of mem­

One reason for the emerging synthesis is that students of the brain and the lnind, after years of going separate ways, have C0111C together [0

ory as I have come to sec it. tlut my goal in writing this book goes beyond describing the ncw

develop an integrated approach that has transformed the study of

synthesis in memory research and relating SOllie of my own discover­

Illemory: cognitive neuroscience. A mere two decades ago, the study

ies and ideas, to include consideration of a puzzle that many of these

of memory was carried out by separate tribes of cognitive psycholo­

findings highlight. Memory, that complex and lIsually reliable asset,

gistS, clinicians, and neuroscientjsl�. Cognitive psychologists studied

can sometimes dcceive liS badly. Yet even though memory can be

memory in the laboratory, but showed scant interest in the world of

highly elusive in sOllle situations and dead wrong in others, it still

memory outside the lab and little or none in the brain. Clinicians-psychologists.

neurologists,

and

forms the foundation for our lllosr strongly held beliefS about our­

psychiatrists­

selves. A head-injury patient I once interviewed who had lost Illany

described fascinating disorders of meTllory, but were unfamiliar with

treasured memories felt that he had also lost his sense of self. He

thl.! elegant techniques llsed by cognitive psychologists to dissect

became so obsessed with the Illissing pages of his past that he could

meillory. Neuroscientists studied memory by removing particular

think or talk of little clse.

parts of animals' brains and then observmg the efe f cts. hardly noticed the findings and ideas of cognitive psychologists or clinicians.

"I can't review my life," he kept telling mc. This important duality-memory's many limitations on the one hand and its pcrvasive influencc on the other--is at the heart of thjs

In the 19805, cognitive psychologists began to emerge from the

book because it is central to understanding how the past shapes the

confines of the laboratory. Some swdied memory in everyday life,

present. I refer to it as fragile power, and it has afft'cted increasing

adding a new richness to their work. Others began to test patients

numbers of us in recent years. An intense controversy has exploded in

with memory disorders, bringing their vast arsenal of experimental

therapy settings. courtrooms, and the popular mcdia as people clailn,

111(rodUClion

Srarchi"g for Ml'lIIory

8

with passionate cOllviction, to have recovered long lost memories of sexual abuse during childhood. Are some of these allc�,.ations based 011 illusory "memories" created, rather than uncovered, in psychotherapy? We have also seen a steady parade of child care workers and others

9

in sOl11e of this reage in deep, elaborative encoding of an event, we arc later likely to

Such dependence implies that we may be oblivious to parts of our

remember that event weU; when we engage in shaUow, superficial

pasts because we fail to encounter hints or cues that trigger dormant

encoding, we will later remember the incident mllch less weU . l t turns

memories. This Illay be one reason why encountering acquaintances

out, however, that superficialJy encoded events can be remembered

we have not seen for years is often such an affecting experience: our

more accurately than deeply encoded events when people are given

old friends provide us with cues and reminders that are difficult to

retrieval cues lhal match exacrly a shallow encoding. Suppose I ask

generate on our OWI1, and that allow us to recollect incidents we

you to think of a word that rhymes with braill. You have carried out a

would ordinarily fail to remember.

64

B u i l d i n g M e lll o r i e s

Searching Jo r AI/ e m o r y

In his relentless quest for self-understanding, the narratOr i n Marcel Proust's novel confronted the disturbing reality that his ability to

65

recount most of his life prior to the operation. but he had great diffi­ culty remembering his ongoing, day-to-day expcrienc{.'S.

recapture the past depended on finding retrieval cues that could

Curiously, however, Neil performed reasonably well at school,

unleash the torrents of memory that he sought. He ultimately came

especially in Eng-lish and m:lthematics. The psychologistS who tested

to realize that he could nOt allow his menta] time travels to depend

his memory wondered how he managed to do so well. To find out,

solely on chance encounters with smells and tastes, and so he instead

chey asked him some questions about an audiotaped book he had

pursued the past by actively seeking out cues and hints that would

been studying, Cider ulilll Rosie, by Laurie Lee. He remembered noth­

help him [0 remember. NJ of us share Proust's problem: [Q understand

ing. Noting Neil's frustration, and realizing that his class performance

bener who we are, we mUSt somehow generate or find cues that aUow

was based on written responses, the examiner asked Neil to write

us [0 remember things that might otherwise remain dormant or sim­

down his answers, beginning with anything that he could recall from

ply fade away. We saw how in the case of patient GR, encollntering

the book. After a while he wrote: "BloodshQ[ Geranium windows

the right retrieval cue (a medical procedure similar (0 one he had had

Cider with Rosie Oranium smell of damp pepper and mushroom­ growth." "What have I written?" he then asked, unable to read his

before) allowed hUll to regain his seemingly lost personal past. We must not, howevcr, confuse these ideas with the norian that all

own handwriting but able to speak normally. The examiner, who was

experiences arc recorded somewhere in our brains, only awaiting the

familiar with the book, immediately recognized that the phrases came

appropriate retrieval cue to be brought il1(O awareness. While con­

direcdy from it� pages,

trolled research has demonstrated over and over that cues and

Intrigued by Neil's ability to write down information that he could

reminders can lead to recall of experiences that have seemingly disap­

nOt express orally, the examim'r asked whether Neil could write any­

peared, it does not necessarily follow that all e:":l'eriences are preserved

thing about incidents related to his hospitalization some two years ear­

and potentially recallable. Sometimes we forget because the right cues

lier, which he had been unable to remember when asked to talk about

are not available, but it is also likely that sometimes we forget because

them. "A man had Gangrene," he wrote, correctly recalling the ailment

the relevant engrams have weakened or become blurred.»

of another man in the ambulance that brought Neil to the hospital.

R_etri('val cues are a bit like the portable metal detectors that scav­

Neil's parents asked him to write down the names of the children

engers sometimes use to try CO recover coins on a beach. If coins are

in his class. He produced a long list. which turned out to be accurate.

hidden somewhere beneath the sands, then the scavenger needs the

When his mother asked him what had happened at school that day,

detector (0 find them. But if no coins remain in the sand, then even the

Neil wrote, "Mum I saw tulips on the way hOllle."This was the first

most powerful detector will turn tip nothing. Our brains include some

time in two years that Neil had been able to relate to his mother a

beaches with hidden coins and others that are barren. Like the scavenger

memory of something that had happened to him in her absence. Neil's parents equipped him with a small notebook, and he began

seeking money, \ve do not know before searching which are which.

to communicate reb'lilarly abom incidents in his everyday life. Yet he remained unable to recount these episodes or:lily. When he wrotc them down, Neil was ullable to read them, and often expressed sur­

NEIL

prise when someone told him what he had written. After an after­

R e t r i ev a l P r o c e s s e s a n d the B r a i n

1988, a

noon's excursion to several r,'lIlliliar locations, Neil was unable to

fourteen-year-old English boy named Neil began radiation

remember anything when asked. But when cold to write down what

trcatment for a tumor hidden deep within the recesses of his brain.

had happened, he provided a succinct. and accurate, summary of the

Neil had been a normal child uncil the expanding tumor began to

afternoon's activities: " We went to the museum, and we had some

interfere with his vision and memory and to create a host of other

pizza. Then we came back, we went onto the Beach and we looked at

mcdical problems. Chemotherapy was eventually successful, but

the sea. Then we came home."

In

eil

suffered heavy cognitive losses. He was virtualJy unable to read and

This case is unprecedented in the annals of psychology, psychiatry,

could no longer Ilame common objects on sight. Neil was able to

or neurology..l6 Neil's tUnlor did damage his brain, including some

66

Dudding Memories

S c a rf/rj"g for M e m o r ),

67

structures that are known (0 be imporr3m for memory. But nothing

that regions of left inferior fromal cortex appear to be particularly

about the condition ofillS brain provides specific clues to how or why

active during encoding processes. In contrast, as Endel Tulving has

he could retrieve recent episodic memories through writing but not

shown, PET studies of memory for words, sentences, faces, and other

speaking.

materials consistently reveal that specific regions toward the front of

There are other indications that the brain uses different systems for

the right frontal cortex are especially active during e"l'licit retrieval of

retrieving written and spoken information. The neuropsychologist

episodic memories-more active than the corresponding left frontal

Alfonso Caramaua has described two patients who suffered strokes in

regions.-tQ Of course, merely observing that a region is active during

differem regions ofche left hemisphere that are usually associated with

retrieval does nor indicate exactly what role it plays in the retrieval

language impairments. Both patients subsequently had special prob­

process. For example, if the right frontal regions that have been :lcti­

lems producing English verbs (they could produce nouns normally).

vated during PET studies constitute a retrieval system that is neces­

One patient had problems writing verbs but not saying them, whereas

sary for recalling all episodic memories, then damage to right frontal

the other had problems saying verbs but nOt writing them.)1

regions should cause a devastating impairment of explicit retrieval;

Caramazza's fmdings still leave us a long way from understanding

such patients should have grdve difficulties recalling past cvenrs. How­

how Neil could recall his recent experiences through writing but not

ever, studies that have included patients with right frontal damage typ­

speaking:. Out these strange cases of disruptions of retrieval raise ques­

ically fail to show generalized impairments of explicit remembering.

tions that are essential to understanding memory: Exactly how does

But these patients can exhibit sOllle memory difficulties, such as prob­

the retrieval process work? What goes on in my mind/brain that

lems in remembermg which of two events occurred more fecencly,

a!Jows the cue "What did YOll do during your sumiller vacation?" to evoke III me the subjective experience of remembering beautiful sun­

and some arc also susceptible to fascinating memory distortions. These . impairments, which I consider later, may be caused by fluiry retrieval

lit days of hiking and swimming at Lake Tahoe? We do not understand

processes.

precisely how the retrieval process works, but some dues art' begin­ ning to emerge.J8

In 1992, neuroscientist Larry Squire and colleagues reported a PET study of explicit retrieval that showed activation of the hippocampus,

One critically important idea is that the brain engages in an act of

which we have already seen plays a role in encoding novel experi­

"construction" during the retrieval process. The idea is well illustrated

el1ces. �l Given that both the hippocampus and the frontal lobes can

by neurologi.st Antonio Oamasio's theory of how the brain remem­

become active when people arc instructed to remember recent events.

bers. As I elaborate in the next chapter, Darnasio and mhers have

it is natural to ask what role each structure plays in the retrieval

argued that there is no single location or area in the brain that COI1-

process. A recent PET study I carried OUt with colleagues suggests that

tains the engr:lI11 of a particular past experience. Posterior regions of

they play rather different roles. Think about what is involved when I

the cortex that are concerned with perceptual analysis hold on to

ask you to recall what you did last Saturday night. Attempting to

fragmentS of sensory experienc(,-bits and pieces of sights and sounds

retrieve th:\t memory involves a r:lir amOllnt of mental work. YOll

from everyday episodes. Various mher regions of the brain, which

engage in search strategies and try to come up with cues. sllch as gen­

Damasio calls convergence zones, comain codcs that bind sensory

erating names of people you might have seen or where you Illight

fragmcnts to olle another and 10 preexisting knowledge, thereby con­

have gone. The mental effort that you expend is an important part of

stituting complex records of past encodings. Oall1asio suggests that

the retrieval process. But there is also the moment when you actually

remembering occurs when signals from convergence zones trigger the

remember the event: you recall that you went to the movies and saw

simultaneolls :lctivatioll of sensory fragments that were once linked

a thriller that kept yOll on the edge of your seat. TillS subjective rcc­

together. The retrieved memory is a temporary cOllStelJation of activ­

ollcctive experience is, of course, an essential pan of the retrieval

ity in several distinct brain regions-a construction with many con­

process. When a brain area becomes active during a memory-retrieval

tributors. :t9

task, it cOllld reAect eicht'r the mental effort associated with trying to

New information reg-drding brain systems involved in memory

remember or the recollective experience associated with successful

retrieval has been provided by PET scanning studies. I noted earlier

remembering. We designed our PET study to discmangle these twO

Se(lrdli llg Jor M e m o r Jl

68

Building M emories

aspects of retrieval. Our results show that increased blood flow in the frontal lobes during explicit retrieval refh:cts primarily the mental effort involved in searching memory. However, silllply trying to remember all event is not enough to activate the hippocampus. Increased blood flow in the hippocampus seems to reflect some aspect of the subjective experience associated with remembering the event.'� In line with cht.'Sc findinb "S, the neuropsychologist Morris Moscovitch proposed that hippocampal and frontal systems may be involved in two different kinds of retrieval processes. One of them, referred to as associalillf retrieval, is an 3moll1aric reminding process that depends on the hippocampus and related medial temporal lobe Structures. Associative retrieval occurs when a cue automatically triggers an expe­ rience of remembering. Everyone is familiar with this sort of experi­ ence: hearing a f.worite song reminds you of where you were when you first heard it; you tell a new acquaintance that she reminds you of an old friend; seeing a cue word in a memory experiment easily

69

system damage leads to severe memory loss for recent experiences, whereas frontal system damage typically does not. When the auto­ matic retrieval system is dysfunctional, the interrogations of the strate­ gic system WIll be fruitless, and it will be difficult to recall recent events. However, if the strategic retrieval process is impaired and the automatic retrieval process is intact. it should still be possible to remember reasonably well in the presence of appropriate reminders or cues. Deficits will be restricted largely to those situations in which effonful, strategic search is required-and this is a reasonably accurate description of what happens in many cases of frontal system damage. The power of modern ncuroimaging techniques raises the possibil­ ity that we may, in rhe nor tOO distant future, learn enough about brain mechanisms of memory retrieval to be able to illuminate the disordlT that rendered Neil mute about experiences that he could describe onjy when a pencil was placed in his hand.

brings to mind the word that was paired with it during the study episode. The other retrieval process, which is referred (0 as

e.fforifll/ or

strategic retrieVtli, involves a slow, deliberate search of memory and depends on regions of prefrontal cortex-perhaps most critically. the right frontal regions that have been activated during explicit retrieval in PET studies.o When I asked you to remember what you did last Saturday night, you relied on the strategic retrieval process to generate hints and cues, which you used in turn to "interrogate" the automatic retrieval process. If the frontal system generates a cue that has a match in mem­ ory. the medial t.emporal system will automatically "spit out" an engram that combines with the cue. Without the aid of the frontal sys­ tem, the medial temporal system must simply "wait" for an appropri­ ate cue to come along and make contact with a stored engram. As we wilJ sec in chapter 3, the medial temporal region works cooperatively with regions toward the rear of the bram where engrams :1Tt� stOred, including areas in the parietal and occipital cortices, forming distrib­ uted networks that aHow liS to encode and recall our recent experi­ ences. Indeed, our PET study TCwaled blood flow increases in specific pam of the occipital and parietal lobes during recall, perhaps reflcct­ ing the dynamic interaction between mcdial temporal srruclUres and cortical regions that results in the construction of a memory. Out we will also see that the medial temporal region is not necessary for retrieving experiences from the distant past. These ideas help to make sense of the fact that medial temporal

C O N S T RU C T I N G M E M O R I E S T h e Role o f t h e R e t r i eval Environ m e n t Findings and ideas concerning brain mechanisms of retrieval are absolutely crucial to undtTstanding memory's fragile power. But it is still important to develop an adequate conceptualization of retrieval at the psychological level. How are we to think about what is retrieved when we recall a past experience? Does the act of retTleval simply serve to activate, or bring into conscious a\varencss, a dormant memory? Suppose, for example, that I provide a retrieval cue such as "tell me about last year's Thanksgiving dinner:' It Illay take you a few seconds to recollect where it occurred and who was there, but by the timc you reach the end of this sentence there is a good chance that YOll will recall some of the basic information. How did this subjective experi­ ence of remembering come about? The simplest account is that the clle somehow activated a dormant engram of the event, and that your subjective experience of fCmcmbering the event, however mcom­ plete, is a straightforward reAcaion of the information that had been quiescent in your mind: a lightbulb that had been turned off i s sud­ denly turned 011. But memory retrieval is nOt so simple. I have already suggested an alternative possibility, rooted in

eisser's analogy that retrieving a

memory is like reconstructing a dinosaur from fragments of bone. For

13 u i l d i u g M e m o r i e s

S e a rcll i tlg for iHe m o r y

70

71

the paleontologist, the bone chips that are recovered on an archeo­

that people reported contained little infonnation about the event they

logical dig and the dinosaur that is ultimately reconstructed from

were trying to recall (the speaker's tone of voice) but were greatly

not the same thing; the full-blown dinosaur IS constructed

influenced by the properties of the retrieval clie that we gave them

them

3TC

by combining the bone chips with ocher available fragments, in accor­

(the positive or negative facial expression) .44

dance with gcncr:d knowledge of how the complete dinosaur should

There have been few other attempts to examine how the proper­

appear. Similarly, for the rememberer, rhe cngranl (the stored frag­

ties of a retrieval cue contribute to our subj ective experiences of

mems of an episode) and the memory (the subjective experience of

remembering. This is likely attributable to what Tulving has called the

recollecting a pasl event) are not the same thing. The stored fragments

"overpowering influence" of the traditional theory that a memory is

contribute to the consciolls experience of remembering, but they are

simply an activated engr." When

he asked what I}art of town she lived in, " 1 felt panic . wash over me anew as I searched my memory and found It blank ' ." �'

SIX

It is excruciating for family members to wimess a patient's p:tinful

�

awareness of his own vanishing mcmory. " Alzheimer's is the cruelc t . of diseases" acknowledges the writcr Glenn Collins, reAecung on 1m _ f.'Hher's str ggle with :I ruthlessly advancing d eJlle' tia. "What s bad is . _ the meanness of it. The knowledge of forgemng. I he frustration and

�l

�

�

confusion and sh:tme of forgening."�l As cognitive function deteriorates over a period of years, mosl

�

�

Alzheimer's patients become progressively les aware of the extent, r . . even the existence, of their deficit... r�ecent eVldencc mdlcates . that thiS

T H E H I D D E N WOR L D O F I M P L I C I T M E M O RY

�

dimming of awareness in Alzheimer's patients is accoll1�alli d by an increase in conf.'lbllbtions on memory tests and by a decline III frontal lobe function. However, signs of frontal dysfunction arc probably 110t , . a necessary precondition for unawareness of deficit: some AlzhclIners

scn�

parient that people did not remember seeing earlier as for

Why did amnesic patients do so well when given letter cues as hints

words they did remember seeing. The resules pushed us toward a

for recently swdied words? If these cues tapped into some sort of non­

strong, seemingly un:lvoidable conclusion: priming occurs indepen­

con�ciolls memory that is preserved in amnesic pati!.!nts, shouldn't it

dent of conscious memory,7

be possible to uncover something similar in peopl!.! without amnesia?

These findings hit us with the force of an avalanche. We behewd

We designed an !.!xperimenr to find out. Our reasoning was simple:

that \w had been able to get a handle on the peculiar kind of mem­

if lener cues tap into a fOflll of memory that is sp ared in amnesic

ory that Warrington :lnd Weiskrantz had documCl1led in amnesic

patients, then we might be able to elicit sllch memory in healthy vol­

patients with th!.! letter cueing task. This "other" kind of memory

lInteers by giving them letters from a previously studied word and

seemed to be lurking in the minds of healthy adults, and could be

asking them to try to guess the answer. Weiskrantz had observed that

tapped by giving the word fragmenr-completion test. We felt a bit like

amnesic patients treat the letter cue test as a gues.�ing game, If young

astronomers must feci when discovenng :l new star or an entire galaxy

adults could also be induced [Q treat the test as a guessing game, we

whose existence had been only suspected: a whole new world of pos­

reasoned, then they might rely on the same kind of memory that War­

sibilities is suddenly open for exploration. I also started to notice manifestations of priming in everyday life. It

rington and Weiskr:lIltz had observed in amnesics.

\Ve carried out our experiment in the summer of 1980. For you

is likely involved in instances of unintentional plagi:trism. Probably the

to get a feel fm our procedure, you should study each of the follow-

best-known case in recent dec:ldes involved the former I3e:ttle George

OC/OpIiS. aI!ocado,

mystery,

Harrison and his 1 970s hit "My Sweet Lord." Unfortunately for Har­

slll'rm; and dimare, Now imagine that you go about your business for

rison, his melody nearly duplic:lted the tlille of a 1962 classic by The

an hour and then return to take a couple of tests. First I show YOll a

Chiffons, "He's So Fine," When a lawsuit was brought against him,

1I1g words carefully for five seconds: assassin,

series of words and ask whether you remember seeing any of them

Harrison conceded that he had heard "He's So Fine" prior to writing

on the earlier list: IwiliglH, assassill, dilloslHlr, and mystery. Presumably

"My Sweet Lord," but denied that he had ilHentionally borrowed

you had little difficulty here.

from the earlier song. Reasoning that the resemblance between the

ext I [ell you that I am going to show

you some words with missing letters. Your job is [Q fill in the blanks

two was simply too strong to be {he product of coincidence, the trial

as best you can: ch----nk. o-t--us, --og--y- - - , -I-Ill-te. You

judgc "held that Harrison's work did infringe through what the courtS

prob:tbly had a hard time coming up with a correct answer for twO

felt rnust have been unintcntional copying of what wa.� in Harrison's subconscious nlcmory." 8

of the word fragments (chipmllllk and bogeymall). But with the other two fragments, the correct answers probably jumped out at you. The

You Illay have enCOlintered instances of this kind of printing, too.

reason thcse fra.gments arc so ea�y to complete. of course, is that you

You propose an idea to a fellow employee or a friend, who seems

just sa\\ the words octopllS and dimate in our study list. This kind of

unimpressed by it or even rejectS it altogether. Weeks or months later,

memory is called prilllillg: seemg the words on the list seems to prime

that person excitedly relates your idea as if he had just come up with

your ability to come up with the correct solution when you try to

it. \Vhen you dra.w this inconvenient fact to his attention-with an . edge in your voice betraying exasperation-you may be f1ced with

complete a word fra.gment. We tested people either one hour or one week after they studied

either heated denial or a sheepish apology born of a sudden dose of

168

S e a r( h i fJ�1? for il'lemory

T h e H J d d e n Wo rl d of i m p l i c i t M e m o r y

explicit memory. A n incident from Sigmund Freud"s life clearly illus­ trates this. Freud had lnaintaim:d for years an intense and tUl11uhuous friendship with the Derlin physician Wilhelm Flie�. He frequently confided his latest ideas and insights to Fliess, and was emotionally dependent on his approval of them. When Freud announced to li(..'Ss a momentOuS new insight-that every person is fundamentally bIsex­ ual-he fully expected Hess to be amazed by the idea. Instead, Fliess responded by reminding Freud that he himself had made � xactly the same discovery two years earlier and told Freud alJ about It, and that Freud had rejected the idea. Freud eventually explicitly remembered the earlier incident, commenting that " [i]t is painful to h:lVe to surren­ der one's originality this way." Inspired by such observations, psychol­ ogists have recently been ablt, to demonstrate a kind of unintentional plagiarism in the bboratOry and tic it directly to prinung.9 R.esearch into priming exploded during the early 1 9�Os, as provocative new articles appeared in scientific joun� als. Priming occurred on ;1 variety of tests in which people were IIlstructed to identify a briefly flashed word or object, or guess an answer, r:lther than try explicitly to remember a word or an object from a list they had studied earlier. For example, Larry Jacoby and Mark Dallas found similar amounts of priming after deep encoding (focusing on a word's meanings and associations) and shallow encoding (focusing on rhe illdividual lcners in the word)--a remarkable result, since deep encod­ ing yields Illllch higher levels of explicit memory than shallow encod­ ing. Yet the priming efTect could be easily elinun:ned. If people heard the target words on an audiotape dunug the study task but did not see a printed version of them, little or no primmg was observed on a later visual test. Something about perceiving the actual word form was cru­ cial for priming to occur.IO Considered together with the results of Ollr word fragmelH­ completion experimcnt, these findinb"S indicated that the new and mysterious phenomenon of priming obeys differcnt rules than the kind of memory th:lt researrhcrs had been investigating: for years. I t bt.!came increasingly clear that part of the Inystery could b e traced to the instructions people are given when their memories are tt.!sted. For example, when :nnnesic paticnt" arc given word beginnings or other cues, and arc instructed to think back to the study list to try to remember target words, they perform quite poorly. But when given the same test cues with instructions to guess or to provide the first word that pops to mind, they do just as well as people without mem­ ory problems. Likewise. depth of encoding inRuences later retention

�

169

when normaJ volunteers try to remember tht, target items, yet has linJe effect when the)' respond with the first word that comes ro mind.ll Scientists love a good mystery, and many researchers tried to figure ou( what printing effectS might mean. Tulving and I had aln:ady staked Out a position: because priming seemed unrelated to conscious recol­ lection, we reasoned that it does not depend on the episodic memory system that allows us to recollect specific incidents &om the past.That system plays a key role in much of what I have discussed in the book so far: remembering what happened at last year's Thanksgiving dinner, remembering where you hit a tee shot during a round of golf, or remembering that you saw the word octopus in a study list. Amnesic patients have little or no episodic memory, but they often show nor­ mal priming. We concluded that the source of priming must lie out­ side the episodic system. But where? Semantic memory-the intricate network of concepts, associations, and f.1cts that constitutes our general knowledge of the world­ seemed a reasonable place LO look. When an anmt'sic patient such as Mickey learns that the first game of baseball was played in Hoboken but does not remember the episode in which he acquired that f.1ct, semantic memory may be responsible. Likewi in Ann's family

culty recalling some aspects of their abuse. As I pointed out in the last

life: her mother admitted that she had withdrawn from Ann emotion­

chapter, however, abused wOlllen with reduced hippocampal volumes

ally during a separation from her husband when Ann was young. "It

alw:_ys remcmbered their abuse and f:1iled to exhibit IIlclllory deficits

was a traditional family," said the fum's narrator. " Every Christmas W:15

on laboratory teSts. And no direct cause-and-effect relationsh.ip between

celebrated. No birthday was ever forgonen and there wcn� always gifts.

sexual abuse and reduced hippocampal volume has been shown. Yet

Hut Ann remembers only anger and a yearning for her mother." Both

even ifreduced hippocampal volume is linked [Q some sort of memory

parentS seemed bewildered by thc accusations of sexual and ritualistic

problem in abused wOlllen, none of these findings proves or implies the

abuse, which included allCb"dtions that Ann's mothcr had attached elec­

existence of a special repression mechanism that aUow·s people to blot

trodes to Ann s genitals and that her father had abused her with tools

out repeated, horrendous abuse immediately or soon after it happens

from a hardware store. Her father, on (he verge of tears as he described

(nor do they explain how memories could be recovered). Cases such as

" the enormity of the absurdity" of Ann's abuse memories, related how

the one reported by Schooler indicate that some people forget multiple

her medical records showed no effects or the viciolls ritual abuse that

inCidents of abuse, but the evidence is not strong enough to warrant any

she recalled, and her school records revealed pcrfect attendance despite

definitive claims abollt massive repression, as opposed to more benign

the torture that had supposedly becn inflicted on the young girl.

forms of conscious suppression and lack of rehearsaL Dissociation, rather than repression, might be responsible for exten­

I f Ann had been dissociating th.roughout childhood to cope with nightmarish riUlal abuse-thus explaining her amnesia for the abuse

sive amnesia in abuse survivors. Dissociation refers to a f:1ilure to inte­

prior to therapy-there should have bcen teUtale signs of a dissocia­

grate different aspeCls of an experience, with the I"(.'sult that it is difficult

tive disorder, such as Spotty attendance at school, serious childhood

to explicitly remember the experience. I have discussed evidence

behavior problems, and the like. " I don't care if it's true," asserted Ann's

that people suffering from dissociative disorders, including multiple­

therapist, Douglas Sawin. " What's important to me is that I hear the

personality patients, can forget large chunks of their pastS. Williams

child's truth, the patient's truth. That's what's important. What ;lctually

suggestS that the women in her study who reported forgetting inune­

happencd is irrelevant to me." Asked abolLt the possibility that a

diately about their abuse may have dissociated the episode. It seems

client's report is a delusion, Sawin did not flinch:" Wc all livc in a delu­

plausible to me that some survivors of sexual abuse Illight repeatedly

sion, just more or less delusionary." If these chilling beliefS are shared

dissociate during recurring incidcms of abuse, perhaps even creating

by other therapists, then it is hardly surprising that there are so few

imaginary identities to handle the abuse. However, recent evidence

well-corroborated cases of recovered memories and so little direct

suggestS that traumatized survivors of sexual abuse may have great dif­

evidence for the operation of dissociation in ther.lpy patients who

ficulty forgetting about abuse-related information, at least when it is

remember previously forgotten abuse..!!!

presented to them in the laboratory. Moreover, if people become

In the Eileen Frank.lin case, the prosecLltion apparently wished to

skilled enough at dissociation to develop total amnesia for traumatic

avoid applying the terlll c/isscx:jalioll to Eileen, probably because it

experiences. it would imply the existence of a dissociative disorder-a

implied the existence of a severe dissociative disorder that could not

seriolls matter. If they have engaged in extensive dissociation, then

be dClllonstrated.1'l I suspect that dissociation may indeed occur in

p:ltients who recover previously forgonen memories involving years of

some cases of genuine forgetting of extensivc sexual abuse. 13m when

horrific abuse should also have a documented history of severe pathol­

it does, there should be a long and well-marked trail of associated

ogy that indicates a long-standing dissociative disol"(ler.17

problems aud pathologies.

i ustrates the point was shown in A particularly affecting case that ll

My reading of the evidence concerning memory for sexual trauma

Ofra Bikers " Divided Memories." A young woman named Ann

pointS toward three conclusions. First, there is no question that some

described how she recovered in therapy memories of terrible 5.1tanic

survivors of childhood sexual abuse forget about single abusive inci­

ritual abuse at the hands of her parents, and also discovered that she

dentS, and some evidence that they may forget multiple episodes of

T h e M e m o r y Wars

S c a rcidug fo r M e m o r y

abuse. This forgetting i s most likely anribmable to some combination of normal processes of memory decay and interference, conscious suppression and lack of rehearsal, and perhaps physiological changes caused by sexual abuse. Second, there is as yet little or no scientifically credible evidence thal people who have suffered years of violent or horrific abuse after the years of infancy and early childhood can immediately and indefinitely forget about the abuse. If convincing evidence of this kind does surface, I believe it will occur in the con­ text of a dissociative disorder. Third, the idea that forgening in abuse survivors is caused by a special repression mechanism-something more powerful than conscious suppression-is still without a scientific basis.

265

after mllch work and searching in psychoan:llysis. For more than fifty years there has been a large clinical literature on psychogenic amne­

�

sias in.dica ng chat events that could nOt be recalled during the anmeslC episode are subsequently recovered. Decause therapists who �eek out forgotten memories of sexual trauma believe that accurate recovery can and docs occur, it seems only reasonable to expect that studies have provided solid evidence that recovered memories are generally accurate. Unfortunately, little . such eVIdence exists. This may be because acmai abuse often OCClirs in

�

se recy, making it difficult to find witnesses or other corroborating eVidence, and also because perpetrators typically deny abuse when it has occurred. In addition to the cases of Ross Cheit, Frank Fitzpatrick, and JR, all of whom recovered memories that were corroborated, another case

R E C OV E R E D M E M O R I E S H o w A c c u ra t e A r e They? Forgetting abom abuse need not necessarily mean that i t can be remembered abrain years or decades later. There are cOllntless banal episodes in our lives that we arc now unable to remember, and Illay never remember. Some lIlay come back to us when we encounter cues dnt trigger what we felt or thought during the experience, but some engrams may have faded away to the point where no clle can elicit them, perhaps because much time has passed and we have nOt thought or talked abom the incident since it happened. Probably the best-known recent example of a recovered traumatic memory is that of Eileen Franklin. Eileen's memory was compelling enough to convince a jury that her father was guilty of murder and [0 convince a psychiatrist as experienced as Lenore Terr that it is genuine. OUt there was no independent corroboration of Eileen's memory. As the attorney Harry MacLean describes clearly in his authoritative accoulH of the Franklin trial, there are reasons to doubt the veracity of Eileen Franklin's recollection of what happened on that tragic day. The recent reversal of George FrankJin's conviction may provide another opportunity to examine the credibility of Eileen's memory in the courrroom:JO Although the Franklin trial received enormous publicity because it was the first time :I recovered memory Iud ever been the ba�is for a criminal prosecution, the idea that a forgotten traumatic memory can be later recovered has a long history. Standard Freudian lore had it that repressed trauTllatic memories can subsequently be recalled, usually

�

wi h solid documentation has been reported by the clinical psychol­ . ogIst Michael Nash. He describes a forry-year-old man who entered therapy in part because he was bothered by an intrusive and unwanted memal image of himself at age tefl, surrounded by a group of threat­ ening young boys. The patient suspected that the image alluded to a sexual e:o.:perience and eventually proceeded to recover a traumatic sexual memory involving the boys. He then contacted a cousin whom, he believed, had been present during the episode. The cousin recalled the incident quite clearly and with considerable embarrass­ ment: he had never forgotten that the patient had been unwiUingly drawn into the group's sexual activities.ll Additional examples of corroborated cases are found among the twelve women in Linda Meyer Williams's recent study who had temporarily forgotten and later recovered memories of docu� mented abuse. Interestingly, none of the women that Williams describes recovered their memories in therapy or used special tech­

�

niql es such as hypnOSIs to hunt for them. Most were spontaneously relllmded of the abuse by cues: Mary remembered when she encou ntered a man who looked like her perpetr:ltor and then started having nightmares; Kim began to remember when someone asked her whether she had been sexually abused; Tanya suddenly recalled her molestation as she watched a movie about childhood sexual abuse.ll Bue if recovered memories are somerimes accurate, we don't yet know how accurate they arc. The psychiatrist Bessel van der Kolk has speculated that memories of temporarily forgotten traumas might be exceptionally accurate. Ordinary experiences that we mull over and

266

Sl'flHh illg Jar M e m o r y

T h e M e m o r y Wars

267

discuss with others can be changed by the retelling; repressed memo­

describe three patients who developed vivid, intrusive images of

ries oftraUI1l3 may remain frozen in their originaJ form:"Conccivably,

highJy disturbing incidents. One patient ;'saw herself as a child with

tr'Jumatic memories then could emerge, not in the distOrted form of

her father about to stab her, and then herself siuing in a pool of

ordinary recall but as affect states, som3tic sensations, or visual images

blood ."� In all three cases, the images were interpreted as flashbacks

(for example, nightmares or flashbacks) that arc timeless and ullmod­

of repressed childhood trauma, and the patients were referred for

ified by further experiences." And, indeed, van deT Kolk and his col­

appropriate psychotherapy. Something was amiss with these patients,

leagues report chat people who have experienced severe traumas

however. An three of them compulsively engaged in unusual rituals,

reexperience them as isolated pictures or bodily sensations accolllpa­

sllch as cleaning or washing over and over again. These disturbed

nicd by intense feeling. whcn:as the same people recall personally sig­

behaviors are characteristic of the psychiatric disturbance known as

nificant (bur nonrraum3tic) experiences in a more storylike narrative

obsessive-compulsive disorder. When the patients were given drugs

for1l1.)J

that arc ordinarily used to treat this debilitating condition, the intru­

We have already seen lhal thc amygdala and stress-related hor­

sive imagery disappeared completely. It turned our that the images

mones play a special role in emotional memories, so it makes sense

wae not flashbacks of actual events; they wen: symptoms of the

that mcmory for trauma does differ

patients' obsessive-compulsive disorder. This in turn suggests that the

111

important ways frolll ordinary

memory. But these differences between traumatic :md nontr:1ulIlatic

diagnosis of repressed childhood trauma was incorrect. Yet had i t not

recollections do not demonstrate or imply that traumatic memories

been for the dramatic effects of the drug, the patients would have been

that have been repressed and latcr recovered are also especially accu­

engaged in a needless-and possibly disastrous-search for repressed

rate. The idea of an unchanging imprint of exactly wh:1t happened at

traumanc InemOTICs.

the time of a trauma brings us perilously close to the dubious notion

The current state of scientifIC evidence concerning the accuracy of

that memory (or at least traUlllalic memory) is likt" a camcorder, pre­

recovered memories of childhood sexual abuse call be summarized

serving :111 aspects of an episode.oW We have seen that this idea is fun­

easily: there are a few well-documented cases, but little scientifically

damentally misguided when applied to ordinary experiences, and I

credible information is available. The COllrts have recently started to

pointed out in chapter 7 that it does not work well for traumatic memories that people always remember. I t would be surprising, cven

ew Hampshire, grapple with this point. In May 1 995 . a judge in . f"lced with nvo cases in which recovered memories formed the basis

extraordinary. if it were to apply to traumatic experiences that arc

of sexual assault charges, brought in experts on both sides of the issue.

buried and then recovered years later. There is currently no scientifi­

"The COlirt finds," wrote Judge William J. Groff in the opening sen­

cally credible evidence to suppOrt the idea.

tence of his opinion, "that the testimony of the victims as to their

The flashbacks of war veterans and others suffering from post­

memory of the assaults shall not be admitted at trial because the phe­

traumatic Stress disorders are sometimes cited as t"vidence for the

nomenon of memory repression, and the process of therapy lIsed in

accuracy of recovered traumatic memories. But we saw in chapter 7

these cases to recover the memories. have not gained acceptance in

that flashback memories often involve a mixture of memory and

the field of psychology, and are 1101 scielltifically reliable." In a more

fanrasy. Flashbacks are heavily influenced by expectations, beliefs, and fears. The contents of :l flashback may say more about what a

recent New Hampshire case, Judge Linda Dalianas cited the same . expert testimony as Groff, but. ruled in f"lVor of allowing the alleged

person believes or fears about the past than about what actually hap­

victim to testifY and allowing experts to testify cOllcerning the gen­

pencd.

eral phenomenon of traumat.ic amnesia. Nonetheless, she also con­

Flashbacks are especially relevant to recovered ITu.!nlories of sexual

cluded that "[tJhe Court will not allow expert evidence regarding

abuse because some trauma therapists say that memories of abuse

either the process or the plausibility of 'recovering' an allegedly

often return as isolated flashback images.lS A recem report by the psy­

repressed memory, because the experts have not offered any data

chiatrists Joseph Lipinski and Harrison Pope dramatically illustrates

either supporting or refuting any theory of how or whether a 'lost'

that such flashbacks mllst be viewed with a great deal of caution. They

memory might be recovered.")?

268

T h e M e r ll o r y Wars

Sear{l!j"g Jor M e m ()ry

I L L U S O RY M E M O RI E S O F S E XU A L A B U S E What Is the Evidence?

269

ory of ritual abuse. But even though thousands of patients have "remembered" ritual acts, not a singk such case has ever been docu­ nlcnte..:d in the United States despite extensive investigative efforts by

We have already seen that some recovered memories are accurate.

state and fednal law enforcement. FI3I agent Ken L:I!Hlmg has inves­

That still leaves a separate and crucial question: Is the..:rc any evidence

tigated over

that people can come to believe that they were sexually abused when

ing corroborating evidence for a single one. A recent report from the

300 cases of satanic cult abuse, for e..:xample, without find­

they were..:n't.? In the sununer of 1 987. Diana Halbrooks's therapist sug­

National Center for Child Abuse surveyed several thousand profes­

gested terminating treatment. " I panicked, became very anxious and

sionals about satanic ritual abuse and failed to turn up conclusive evi­

increasingly depressed," DIana remembered. "I felt it was due to think­

dence for tillS kind of abuse or for the organized, intergenerational

ing about not seeing hlln regularly any longer, bur he informed me

cults that have been implicated in recovered-memory cases.l8

that it was not thac at all. that it was due to abandonment pain related

These failures to document cult abuse do not necessanly mean that

to my father." She became increasingly depressed. but despite her best

no satanic cult� exist or that no ritual ;lbuse has ever occurred. Events

effort'>, "I couldn't come up with anything that my father had done to . cause such pain." In a last effort to find out what her f'Hher had done,

such as the 1995 Oklahoma City bombing, the..: cult-rdated nerve gas

her therapist asked Diana to write down whatever came to Illllld

by the Wisconsin murderer Jeffrey Dahmer provide painful reminders

while she was in a hypnotic trance, telling her that "writing in a trance

that people arc capable of terrible deeds.J9The human capacity for evil

with my eyes dosed would aUow my unconscious co speak freely."

is not at i�sue, but the..: human ability to develop amnesia for repeated

Diana had become adept at ente..:ring hypnotic trance states in therapy.

acts ofbrutaliry IS very much in doubt. Until convincmg evidence

When she opened her eyes this rime, she was met with a shocking

forthcoming,

sight: " 1 had written that my father molested me."

ritualistic horrors perpetrated by cults arc based on illusory recollec­

With a door to her unconscious seemingly opened, Diana contin­

attack on a Tokyo subway, and the horrific acts of bur chery carried out

is 1 conclude that most, if not all, recovered memories of

tions. And since most recovered memories of satanic ritual abuse

ued this exercise. At the same time, she entered a weekly support

emerge only when therapy has begun, these cases lend support to the

group that her therapist had initiated, consisting of wome..:n who were

idea that false memories of brutal traumas can be created during ther­

abio explonng recovered traumatic memories. The atmosphere of the

apy.

group was highly charged, as women discusse..:d and sometimes acted

Because Diana Halbrooks's ntual abuse memories seemed so out­

OUt their memories and dreams of terrible events that had come back

landish, her doubts about the reality of these and her other recovered

to them in therapy. Diana recalls: "The memories that others were

recollections continued to grow. But these doubts met resistance from

sharmg in the group were getting more and more bIzarre: satanic rit­

the people in her support group and her therapist." I continually ques­

ual abuse, babies being sacrificed, group sex and horrible tortures." It

tioned the memorie..:s, doubted them," Diana acknowledges, "but

was not long before the same kinds of horrific incidents began show­

when I questioned the therapist, he would yell at mc, tell me I wasn't

ing up in Diana's OWll trance writings. "In November of 1988, I wrote

giving my 'little girl within' the benefit of the doubt. Tell me that I

in a trance and 'recalled' the first memory of satanic ritual abuse.

was in denial.

Everything just seemed to go downhill from there." By

1 989 Diana

1 didn't know what to believe. But I trusted lum." Even­

tually, most members of the support group recovered memories of ritual abuse..: and almost all, including Diana, were diagnosed as l11ulti­

had renlelnbered ki!ling a baby. A surprisingly large number of recovered-memory cases involve satanic ritual abuse, including reports of child sacrifice, cannibalism, and various gory rituals. A survey of members of the American Psy­

pic personalities. J)l:ma escaped this toxic therapy and has managed to reassemble the pieces of her life. She no longer believes

du[ her recovered memories

1 2 percent reported having

have any basis in realiry. In the world of recovered memones, people

treated ritual abuse cases. Memories of satanic ritual abuse nearly

like Diana arc called relWCfors. Retractors live Illost of their lives with­

always emerge during therapy; clinicians who treat these patients

out any memories of abuse, procced to recover mcmories at some

acknowledge that it IS rare for them to emer therapy with any melll-

poine, and then later come to believe that those memories are inac-

chological Association n::waled that

The M e m ory

Se,u(/I i/l,{! fo r M e m o r y

270

Wan

271

CUraH!. A growing number of people are retracting their memories,

tory can be implanted in adults, but the evidence is certainly consis­

but that does not necessarily mean that all their memories are illusory. . A person might retract a memory because of pressure from f'unily or

tent with that possibility.')

much to bear. Hut when the memories arc as improb:lble as those that

Hypnosis played a key role in Diana Halbrooks's story as well. She . 1ever remembered any sexual abuse by her f 1ther until her therapist � msrfllcted her to engagl.! in automatic writing dllring a hypnotic

Diana Halbrooks recovered, the most reasonable interpretation is that

trance. I know of no scientific evidence that writing down whatever

friends or because the pain associated with the memory is simply roo

comes to mind during a hypnotic t.rance promotes accurate recall of

the events do not have any basis in reality. Dian:1 shares features in common with other retracrors. A recem survey of twenty women who retracted their recovered memories of sexual abuse revealed some striking silllliariries among thcm ..oo

lIle­

teen recovered their memories during therapy, and all of them report

forgott�n experiences. Trance writing was popular over a century ago, when It was used in seances as a tool for psychic communications.

�cattered anecdotal observations suggest that trance writing lIlay occa­ slOnalJy lead people to produce implicit memories of long forgotten

that their therapists inAuenccd the development of their memories.

events that seem strange and unfamiliar to them.13ur it is impossible to

Only one of the twemy did not participate in therapy. This woman

tell whether something th,lt pops to mind during trance writing is an

to Helll, the bible

accurate mcmory of a distant event or a reflr.:ction of current concr.:rns

of the recovery movement that has been roundly criticized for

and fears-just as there is no way to tell (without extnnal corrobora­

admonishing people to believe that they were abused even when they (,1il ro remember it. Nearly all of the retractors-90 pcrcem­

tion) whether memories recovered with the aid of hypnosis are true or . f1lse. With all the talk of satanic ritual abuse in her weekly suppOrt

repofted that some son of trance induC(ion was used in therapy to

group, i t is hardly surprising that a ritual abuse "memory'· eventually

recover IllCIllOTles. Hypnosis was the

popped to mind when Diana cng:tged in tranc(.' writing.«

recovered her memories after reading n,c

reported by

85

Gmmge

most common technique,

percent of the women. Trance writing, rcgression, and

Hypnosis stands Out as a common denominator among the s.1mple

suggestions of abuse were also widely reported. The majority of the

of retractors, but sharing memories

women (70 percent) said that group therapy inAuenced their recov­

important tool used by therapists who believe that it is important to

ered mcmories. "The group progressed from eating disorders to child­

hunt for repressed memories of sexual abuse.'1 Discussing traumatic

hood sexual abuse, to incest, co S R A Isatanic ritual abuse!," recounted

recollections with other abuse survivors would no doubt be reassuring

III

a support group is also an

one retr;ICtor. "Eight out of ten members developed SRA mcmories,

and helpful when those recollections are real; the experienccs of Diana

the tWO who didn't were cold they were in denial." Another wOlllan

and Other retractors teSlify to the power of a group in also helping to

commented that " [ilfyou don't have a memory yOll feel like you have

shape and maintain memories of experiences that never occurred.

to come lip with one to compete with everybody."" Obviously, there can be no unequivocal experimental data showing that :111 illusory memory of a sexual trauma can be implanted, but a

Social psychologists have for decades documcnted and discussed simi­ lar kinds of social inAuences, although there has been surprisingly lit­ tle work concerned specifically with social influence on memory.....

clever attempt to surmount this problem in a related area was reported

Guided imagery. or visualization, IS another mr.:thod recommended

by Nicholas Spanos, a hypnosis researcher. Spanos and his colleagues

by various practitionr.:rs to retrieve repressed 1ll1.!1lI0ries. Here, patients

conducted an experiment using a variant of the hypnotic age-regres­

are encouraged to imagine abusive incidents til:\{ they seek to remem­

sion tl.!chnique, in which people are given suggestions to ';regress" to

ber, attempting to create pictures in their minds of what Jlljght have

a very young age. His subjects were given suggestions to "regress" co

happened. When sexual trauma has actually occurred, guided imagery

a past life. Roughly half of them came to believe that they had indeed

can be a usdili therapeuric technique. Studies by the psychologist Edna

lived a past life. And when Spanos suggested to some that they had

Foa and her cOUeab'lll'"S have shown that the imaginary reliving of a rape

suffered abuse as children in their past lives, they developed more

produces a significant reduction in symptoms of post-fraUlnatic stress

';melllories" of the abuse than people who had not becll given such

disorder. But it is quite another matter to use !:,'llidcd imagery in an etTort

suggestions.u Because thi� experiment was couched in a "past lives" . contl.!Xl, it does not settle the issue of whether a f1lse sexual-abuse his-

to dredge up a supposedly repressed memory of an event that may or rnay not have taken place. I have already pointed out that freely irnagin-

272

Searcidtlg fo r Memory

ing an event, and then exploring i t and talking about it

T h e M e m o r y Wars as

if i t were real,

is a potentially powerful means of creating the kind of subjective feeling chat accompanies an authentic l'nemory. After enough visualization 3nd discllssion, patients Illay be unable to sort out whether the memory has come to "feel"like a genuine one because the event actually happened or merely because thL')' have been imagining it and talking about it.This idea receives suppon from studies by Ira Hyman and his colleagues that . have documented that using g1.lided imagery increases f'llse memories of childhood events (see chapter 4). Recent PET scanning studies con­ ducted by Stephen Kosslyn and his colleagues have shown that some of the same regions in the occipital lobes are involved in both visual imagery and perception. This Illay be one reason why incidents that people frequently imagine can come to feel like events that actually occurred: imagined events are generated by some of the same neural machinery that contributes to the perception of actual event�"1 Just as there is no good evidence dut techniques such as guided imagery and hypnosis can aid accurate retTieval of distam, forgotten, or repressed memories, there is no hard evidence that these techniques are specifically responsible for the creation of pscudomcmories in therapy. And it is of course possible that

III

cases where real abuse has

occurred and has been forgotten, such procedures might help some patients recover their memories. llut unless a ther:.Ipist can cite evi­ dence that a specific memory-retrieval technique enhances accur:ltc recall without promoting false recollections, it is inappropriate to con­ tinue to use unproven and potentially hazardous memory-retrieval t{�chniques. Nonetheless, a recent survey of 1 4 5 doctoral-level psy­ chotherapists in the United States indicates that close to one-third sometiJ'nes use hypnosis to help clients remelnber child sexual abuse, and about the s.,me percentage report using g1.lided imagery."'" I believe that three major conclusions are warranted concerning the likelihood of therapeutic implantation of false memoriL"s of sexual

tr.uuna. First, there is no conclusive scientific evidence from controlled

273

(;liluR."S to document 5.1tanic ritual abuse; recovery of lIlernori("S for seem­ ingly impossible events

(past lives

and alien abductions); growing num­

bers of ther.tpy patiems who have retracted their memories; the constructive nature of memory for emotional events; and the risky mem­ ory-retrieval techniques advocated by some proponents of recovered 111('1ll0ry therapy.Yet we still lack solid data concerning the prevalance of therapy-mduced pseudomemorit."S. We simply don't yet know whether illusory memories of sex1.1al abuse are exceedingly rare, as some clinicians have claimed, or whether they are widespread, as critics of so-called recovered memory therapy have :llb"ucd. It seems unlikely, however, that they can all be written off to just a handful of w:l)'\v.mi therapisrs.1'I Viewed in a broader historical perspective, the idea that people can acquire convincing but inaccurate memories during therapy should not be all that surprising. The philosopher Ian Hacking points out that over a cenlUry ago, Pierre Janet routinely rr!,!aced traumatized patients by implanting false memories to replace accurate but painful recollections of horrible events. For instance. he hypnotized one patient who was overwhelmed by childhood memories of sleeping next to a girl with an acute skin disease called impetigo on her face, leaving it covered by pus­ tules and crUSlS. Janet replaced the traumatic memory with a pleasant but illusory image of a lovely face. Elizabeth Loftus described something similar in a 1982 paper. nearly a decade prior to the recovered-memo­ ries controversy. In an ironic twist, given later developments, she quoted nvo psychotherapists who implamed entire false histories in p!'!ople as a \vay ofmaking them feel bener. Working with people who had been fat all their lives, they successfully implanted f.'llse childhoods in which the patient� had grown up thin. The therapists noted that they "could very easily install mernories in YOll that related to real world experiences that never occurred:· "Made-up mell10riL"S can change you just as well as the arbitrary perceptions that you m:lde up at the time about 'real world events,' " they reflected. "That happens a lot in therapy."50

rt.'search t.hat f.,lse memories of sexual abme can be created-nor will such evidence ever ex ist, because of ethical coltSiderations. Second, there is likewise no defmiti...'l! scientific evidence showing that therapy per sc or specific sUg&1CSth'C techniques are alone responsible for the creation of inaccurate memories. Third, several separate strands, when considered together, support the conclusion that some therapists h:lve helped to crc­ :lle illusory recollections of sexual abuse: the experimentally documented malleability of memory in response to suggestive influences; evidence that hypnosis call produce compelling but inaccur.tte pseudomemories;

D I S T I N G U I S H I N G A C C U R AT E A N D I L L U S O RY R E C O L LE C T I O N S The Role o f Implicit Memory I f accurate recovered memories and illusory recollections of sexual trauma both exist, then an important question imnH.!diately arises: Are there any scientifically based critcri:l that allow us to distinguish mem­ ories of events that aCfUally occurred from false r!'!collections? The

S e a rchi"g Jor A1efll o r y

T h e M e m o ry Wars

answer, unfortunately, i s no. Laborarory studies have provided sugges­ tive clues about. differences between rnemories of actual events and imagined events, but there is no research that allows ciiniciall.s or sci­ entists to judge unequivocilly rhe historical truth of a traumanc mem­ . ory recovered in therapy Noncthdess. some therapists have offered lem� suggestions about how to distinguish an accurate recovered n .

published in various popular publications for incest survivors and that

from an illusory one.S1 One intriguing possibility involves ll11phclt

Illemory for a past experience, it is essential to demonstrate that a behavior or symptom is specifically related to that experience. With

memory for forgotten traumas.

general symptoms of the kind that are described in popular checklists,

274

,!

. Sigmund Freud and Josef 13reuer's classic studies of hystena described patiems who could not explicitly remember childhood sex­ ual abuse. but experienced disabling fears, nagging anxieties, intrusive

thoughts. or disturbing images that reflected implicit memory for {he trauma. However. these cases proved difficult to interpret because independent corroboration of the event was often lacking. As I noted

275

include such items as low self-esteem, sexual difficulry, eating disor­ ders, depression, fears of abandonment, and so forth. Here I agree with critics who maintain that such symptoms are so general that they could apply to many people.5o< To invoke implicit

it is diffICult to establish causal links in individual cases, allhough it i s o f course possible that they sometimes exist. The psychiatrist Harri­ son Pope and his colleagues found no evidence th:H childhood sexual abuse is a risk factor for the eating disorder known as bulimia, even though bulimia has been often cited as a telltale sign of forgonen abuse.J� We do not yet know whether particular kinds of behaviors

in chapter 4. Freud later abandoned his early belief in the reality of sllch traumatic experiences in favor of the idea that they are often fan­

and symptoms are unconscious remnants of traumatic abuse.

tasy-based false recollections.u Breuer and Freud's early observations are germane to contempo­

enter therapy with unexplained fears of or reactions to a particular sit­

rary controversies because some have contended that true recovered memories-not false recollections-are nearly always preceded by behaviors :lI1d symptoms thlt reflect unconscious or implicit memory

�

for repressed trauma. Lenore Terr's research with traumatized child en shows that such implicit memory effens can occur. She studIed twenty children who had been subjected to various kinds of traumas prior to rhe age of fIVe; in almost all cases, the trauillas were corrobo­ rated by eyewitnesses, police reports, or other means. Terr found that nineteen of the twenty children-including several who could not relllember their traumas in words-showed the influence of the trauma in their play, fears, and other nonverbal behaviors. Terr nOtes that only one child-the single case of false memory in her sample-­ had no behavioral symptoms of trauma. This little girl had heard fam­ ily members talk about the traumatic event but had not actually

:

experienced it. IfTerr's observations apply to other situations, the�1 he

In some cases, however, more specific symptoms exist; patients may uation. smell, or object. "Attraction to, avoidance of. or distress around objects or situations unexplained by your own history arc \varning signs of repressed memories," writes the therapist Renee Frederick5011. ';Ouring

sexual abuse. your mind focuses on the events :md cir­

cumstances surrounding the abuse. You may bury the memory, but you store the reaction to the objects or situations that remind you of the abuse."56 This is a plausible suggestion. In chapter

8

we cncounrered

instances of implicit memory for forgotten traumas in psychogenic amllesia patients, and I suggested a role for til(' amygdala l1! mediating persisting emotional aftereffects of experiences that arc not recollected consciously. Uut it is still a long leap to interpreting a patient's unex­ plained fears, attractions, or dislikes 3S implicit memories of sexual abuse. One problem is that even a specific symptom could have any number of causes. In a repressed memory case cited by Frederickson, a patient felt an unexplained rev ulsion toward forks. Subsequemly, she

presence or absence of implicit memory might indeed help dlslHl­

recovered a memory of her aunt abusing her with a fork. Frederick­

guish between true and false recovered memories o f sexual ablls .1l . . . But there arc problems in trying to apply this Idea to mdlVldual

son concluded that the abusive episode is rhe origin of the patients

cases 111 which memories of long-forgotten traumas are recovered in

for the abuse? Possibly. But suppose Ihat Ihe fork revulsion had noth­

�

therapy. Some therapistS include a vast range of symptoms and behav­ iors as possible indications that a person who has no cxplicit memory of abuse has nevertheless been influenced by it. These symptoms pop­ ulate the notorious " checklists" for possible past abuse that have been

unexplained revulsion toward forks. Couldn't this be implicit melllory ing to do with sexual abuse, yet was one of [he factors that encour­ aged the patient or the therapist to explore the possibility that she had been abused. Or suppose that the therapist focused on the meaning of this revulsion because she assumed that it was a symptolll of abuse.

276

Sellrril i llg for JHclI/ () r y

T h e M e m o ry Wars

Then the symptolll could become a focal poim for constructing an

BEYOND C O NTROVERSY

illusory memory. Once therapist and patient become attached to the

As I write these words the recovere d memories controversy continues t� rage. Yet we stilJ have little good scientific cvidence that bears directly on the key issues that I have discussed. Few times in the his­ tory of psychology or psychiatry has the ratio of data to impassioned argument been so low. If we are to find the truth, I believe that we need to recast the tone of the recov ered-memories debate from its present black and white polarities to one that acknowledges more shades of gray.

idea that the symptom reflecrs forgotten abu�c, it should not be sur­ prising when a patient startS to produce images, thoughrs, and feelings that in some way relate to the- symptom. Nobody can say for sure whether slIch a process was operating in Frederickson's patient. but the psychiatrists Susan McElroy and Paul Keck have.: recently described a case in wl1lch it dearly played a role. Ms. B. sought therapy because of depn.:ssion and intrusive thoughts of harming her infant. She also told the therapist tim as a child she had experienced unwanted images of being raped, and wondered whether her disturbing ideas could be symplOllIs of unremembered sexual abuse: "The therapist responded to Ms. B. that these symptoms wcre 'clear evidence' that Ms. B. had in fact been sexually abused as a child, and instructed her ro draw pictures of anything that came to her mind:' After six months of trying, Ms. B. developed detailed recollec­ tions of sexual abuse involving her sister and brothcr-in-Iaw. However, Ms. B.'s condition did not improve, her sister angrily denied the abuse, and she could find no corroborating evidence that any abuse occurred. Ms. n. finally concluded that her memories were false and sought alternative lherapy. Eventually it was discovered that Ms. B.'s intrusive thoughts were attributable to an obsessive-compulsive disor­ der. As in similar cases I mentioned earlicr, they disappeared when crt.'ated with appropriate medication.\' It seems likely that in some cases of recovered memories, unusual fears, anracrions, and related symptoms Illay well turn out to be implicit nH.!Il'lories of prior abuse. In other cases, like Ms . B.'s, such symptoms may provide a basis for creating false recollections in response to suggcstive probing. rather than reflecting the influence of a traumatic memory that had been there prior to therolpy. Inferring implicit memory is a tricky business that calls for careful comparisons and "ysrcmatic reasoning. Behavio� and symptoms may have Illany possible causes, and it is diffiCllit to say whether one partic­ ular experience is the source of a specific behavior or symprom. In the laboratory, experimenters can control the events that give rise to implicit memories. In therapy settings where patients do not explicitly remember being abused, and where we do nor even know for cenain that the abusive event occurred, it is impossible to make comrolled comparisons. A therapist who engages in undisciplined interpretation of fears, altractions, and other symptoms as signs of implicit memory for forgotten abuse may be taking a stcp down a road to disaster.

277

For one, the notion of ''false memory" irself is too coarse to do jus­ . tJc� to the complex relati. ons between memory and reality. When a pa�lent remembers growing up in an abusive satanic cult that did not eXiSt, then we have a belief that defie s reality. Even if the ritual abuse . memory IS a meraphor for some other distressing experience, the . . . memory IS historically wrong in a way that most memories arc not. But what of a woman who was emo tionally brutalized by a neglect­ ful parent, or perh�ps exposed to sexua lly inappropriate language, . behaVIOr, or fondlIng, and then remembe rs incest when none occurred? he incest memory is iJIuso ry, and should be regarded as such, but It Illay capture something illlportalH about the past that should not be disl lissed. Historical truth can be respected while at d e � . . . sam� tllne domgJustlCe to a patient's narrative truth. We need to rec­ ognIZe that memories do nor exist in one of two states-either true or false-and that the important task is to examine how and in what ways memory corresponds to realit y.� Contrary to wl at some have said, there is a middle ground in the � re�0 ered-llle1l10T1eS debate; the prob � lem is to identify it. I believe that thiS IS Ol r best hope for resolving [he bittcr and divisive arguments � �hat COntllllIe to rage among patients, families, and professionals. Polit­ . Ical posturll1g and grand generalizations on both sides of this debate should come to an end. Risky thera peutic practices lleed to be stopped. Better techl1lques must be developed that allow us to distin­ . gUIsh et�veen accurate recovered memories and illusory memories that anse 1Il re ponse to suggestion . Achi eving � these objectives should . . help to IlIlnUll1ze the possibility that those who were not abused come to embrace the psychologically devastating belief that they were redu�e (and, one hopes, end) false accusations that shatter lives an, allllhes, and als maximize the credibility of the memories reported � . y genUll1e survivors of sexual abuse. Sadly, legiti mate concerns about pseudomemories have probably helpe d create doubts about the accu-

!

�

?

�

!

T h e M e m o r y Wa r s

FIGURE 9 . 2

279

rate recollections of some survivors of actual abuse, an outcome that should not be tolerated by therapists, researchers, or society. When Diana Halbrooks reflects on the peculiar memories that she now understands were illusory, she returns over and over again to a single theme: the importance of her family. Diana appreciatcs that there were problems in her early f.1l11ily life. She did not grow up in tht.: idyllic 1950s family that served as a model for many of the post-World War I I generation . l3ehind the vcneer of smiling faces in old photographs and home 1I10vies, there were real difficulties that Diana needed to confront as an adult. (See fi!:,"Ure 9.2.) Diana Halbrooks came to understand that she could acknowledge a painful reality without vilifying or abandoning her parents. "We are rebuilding," she reflects, "and trying to make up for lost time, making every moment count." Not all the victims of the memory wars will be able to unite with their ("milies again. This realization is sad as well as ironic. Our families serve as social repositories for autobiographical n:collections from many times in our lives; we revisit f:woritc episodes, stories, and momentous occasions during holidays and other family gatherings. Yet for some patients. recovering memories of distant trau­ mas-whether aCCllrate or illusory--serves to disconnect them from

Lorie Novak, "Fragments," Courtesy of the artist.

1987. 16�

x

22". Color photograph.

Novak is concerned with the nature of memories for family life. underscor­

ing thl' discrepancy between our idealized versions of childhood and the more troublesome fecl ings that often lurk beneath the surface. In " Frag­ ments" a Polaroid snapshot of a classic 19505 family that includes the artist as a young girl creatl'S a nostalgic sense of a harmonious past. But the torn and fragmemed images in the foreground show Novak at all older age and imply that the innocent f.1m ily portrait masks a more complex and perhaps painful reality

.

one of their richest sources of personal history. The beginnings of our life stories are written in our families, and when we try to make sense of these stories near the ends of our lives, it is often to the family that we return.

S t o r i es o f E l d e r s

281

is especially concerned with reviewing, and attempting ro understand, a family tragedy that has colored h i s entire life: he killed his younger

TEN

brother in a hunting accident. The in cident was too painful for him ro face when he was younger. Like others who have lived through terri­ ble events, he tried his best for years to avoid the wrenching rnemory, mostly by working endless hOllrs o n his farm:"lf I kept at it, kept that ditchbank so clean you could eat Ollt of it, kept the weeds alit of the tobacco and corn, gOt rhe whole pbce looking better than it ever Iud, m:lybc everybody would forget some day that rd killed my own

STORIES O F ELDERS

brother. Maybe I"d forget, (00."2

He never did forget. Out it was only with the distallce in time pro­ vided by old age that Littlejohn could ponder what had happened. The novel charts his battle to comprehend the haunting memory, and ultimately to plaer.: it in the perspective of the fundamcntally decent

E I G H TY-TWO- Y E A R - O L D Littlejohn McCain, the central character in Howard Owen's novel

Lift/rjo/Hl,

is embroiled in a struggle with

memory. As he approaches the end of his life, this farmer from Nonh Carolina is often plagued by his diminishing ability ro remember his recent experiences. After forgetting frequently to turn off the burners on hi � stove, he becomes so flustered that he mounts a large sign on

the kitchen door with the words TURN OFF BURNERS embla­

zoned on it. Sometimes Littlejohn forgets where he is or what he is doing, as in a jarring episode at the grocery store: " I had JUSt pm twO

�

cans of Camp �U's cream of chicken soup i n my cart and was looking for the self-mlllg Rour when I just blanked OU[. I couldn't quite remember what I was doing there. It had happened once or rwice before, but never this bad, or this public, at least. I looked around, and there ar the end of the aisle was the meat counter. It seemed like I rec­

life that he has led. Littlejohn reminds himself of the broader context in which the awful evem occurred by turning to a symbol of the good things that happened ill his life: a crepe myrtle flower that his wife cul­ tivated when they first moved into their home. The house was later destroyed by fire. but even after selling olT theIr land, Littlejohn saved the largest crepe myrtle and used i t to help create a balanced recol­ lection of the past: "Th:n crepe myrtle comes out every summer all pink and be:lUtiful,just when everything else is dying, and it helps me to remember what a fine life we had, in spite of everything."} We see here a kind of age-related heightening of memory's fragile power: the elderly Littlejohn is increasingly plagued by memory's f.1ilures at the same time that he takes refuge in its cOlllpdlmg force. Pat Potter, an artist from Alabama who has spent a good deal of ti111e listening to the reminisccnee/_ . ogy, 2, 331-3:>0.

�

.•

R�ndall, I�. ScOIl. T. M.. Brunen, R. A.. St'ibyl. J. p.. Sourhw ick. S. M.. R. C ' MeCaNhy. G., Cham.:-)', D. S. . &: imllS, It. U. (1995). MIU-bJ.e.l d ll1t'a_ : sl,rcment of h�ppoc�l1lp;11 volu me 111 pltiems widl combal-re\alt-u POSltDUrl 1alic .I(r.:ss . dlsordcr. Ilmmlim jlllmwi ojPsydri"rrJ� 152, 973-981. Hrcmnn.J. D., Sicinberg. M.. Southwick. S. M.,jo hnson. D. Il.. . &: Ch�rncy. D. S. (1993). Bremner. J. n, Delaney.

stru�Ill rc:?

Use ofdie �111l.iC21 interview fur DSM·IV rllssociali\"e disorders for systelll3tic aSSCSm S em of d l.SSocl� l1\ "e symptoms m posIIDum�lic SIren dIsorder. ,-imnir,iri"", 22, 326-343. Coons. P. M. (1994). Confirnlllion of childhood �hLLSC in chiLd and :adol.:sc",m ca� ofmul­ tiple pgllllion, 18. 43-57. Coril!. I. H. (1907). Thc Lowell ca�e of amnesi:l..jollmill ofAbm'mwl Psydlolgr al Psy� on rn�gnetic �nance ImagIng in chromc post�tr:;rUIlUtlC SIms d

mll�n

••

!,

.•

(

drildence ror adult motor conex pl�fA/utI &1I1r"t: ,\IIII,iJllt ptNolllllily ill AmmClJII (Ulw". Wash­ ington. DC; Snuths.onian Institution Press.

Kidder. T.

(1993). OIJfiitrrds.. New York: Houghton MllHm.

cogniti\,l: ";c\Il. III K. S. Bow_ 149-211). NewYork: Wiley. Kihimulll,j. F. (in press �). Exhumed mC'Ulory. ln S.j. Lynn & N. l� Spanos (Eds.), Tnl//r !i

/01, 749-783. Malldl(Or. G. (1980). Recognition: Thc Judgment of

previOUS

McCldbnd. J. L (1 995). Constructh'e mcmory and memOT)' diMoruons: A par:.t1lel-dimib­

uted processing appr