UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

An architectual mind : the nature of real, created, and fabricated memories for emotional childhood events Porter, Stephen 1998

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1998-346110.pdf [ 15.53MB ]
Metadata
JSON: 831-1.0099363.json
JSON-LD: 831-1.0099363-ld.json
RDF/XML (Pretty): 831-1.0099363-rdf.xml
RDF/JSON: 831-1.0099363-rdf.json
Turtle: 831-1.0099363-turtle.txt
N-Triples: 831-1.0099363-rdf-ntriples.txt
Original Record: 831-1.0099363-source.json
Full Text
831-1.0099363-fulltext.txt
Citation
831-1.0099363.ris

Full Text

A N A R C H I T E C T U R A L M I N D : T H E N A T U R E O F R E A L , C R E A T E D , A N D F A B R I C A T E D M E M O R I E S F O R E M O T I O N A L C H I L D H O O D E V E N T S by S T E P H E N P O R T E R B.Sc. (Hons.), Acad ia University, 1992 M . A . , The University o f Bri t i sh Columbia , 1994 A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F D O C T O R O F P H I L O S O P H Y i n T H E F A C U L T Y O F G R A D U A T E S T U D I E S (Department o f Psychology/Forens ic area) W e accept this thesis as conforming to the. reauired standard . T H E U N I V E R S I T Y O F B R I T I S H ^ L U M B I A August , 1998 © S t e p h e n B . P o r t e r , 1998 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, 1 agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of "f^ljcX^ 6q^j The University of British Columbia Vancouver, Canada Date Auj. JO. 7 7 9 f t DE-6 (2/88) The Architecture of Childhood Memories Abstract The false/ recovered memory debate has highlighted the complexities involved i n assessing the validity o f memories for emotional chi ldhood events. This dissertation begins by tracing the history o f the dominant school o f thought o n memory, the spatial perspective, as wel l as far less conspicuous reconstructive views, and challenges influential modern spatial views (e.g., repression) i n light o f a more defensible reconstructive model . T h e empirical component o f this dissertation was designed to compare the nature o f real, created, and fabricated chi ldhood memories for emotional events wi th in individuals. The critical issues being addressed i n the experiment were: (1) whether people could come to remember false ("created") memories for emotional events; (2) i f so, whether differences existed between created memories and real and /or intentionally l ied about (fabricated) memories, and; (3) whether there were individual differences i n susceptibility to created memories. U s i n g a variation o f an approach developed by H y m a n , Husband, and Bill ings (1995), a questionnaire was forwarded to participants' parents inquiring about six categories o f negative emotional events {serious medical procedure, serious animal attack, getting seriously hurt by another child, serious indoor accident, serious outdoor accident, and getting lost) w h i c h the participant may or may not have experienced between the ages o f 4 and 10 years. Parents were asked to describe each event w h i c h had occurred and to give a number o f specific pieces o f information relating to the event. Based on the questionnaire informat ion (85% response rate), 77 participants were interviewed about each o f a: (1) real event; (2) false event; and (3) fabricated event, i n three weekly-spaced interviews. O v e r the three interviews, the interviewers attempted to implant a created memory for the false emotional event using encouragement, context reinstatement, guided imagery, and instructing daily recall attempts. In the first interview, participant were asked about the real event and the false event (counterbalanced), each introduced as a true event. They were provided the event tide and four specific pieces o f informat ion to cue their memories (their age, location, season/ m o n t h , and people The Architecture of Childhood Memories tii present), based o n questionnaire information (contrived for the false events). In the second interview, participants were re-interviewed about their memories for the false event fo l lowed by the implantation procedure. In the third interview, participants were again interviewed about the false event w i t h the same interview approach. Finally (after the last attempt at recalling the false event), they were provided written instructions to fabricate a ch i ldhood memory, again w i t h an event category and four informat ion clues, given preparation rime and a monetary incentive, and interviewed about the fabricated event w i t h the same format as the other two memory types. F o l l o w i n g transcription o f the two or three (if a created memory had emerged) final memory reports, the memories were compared on several dependent measures, collectively designated the M e m o r y Assessment Procedure ( M A P ) , relating to their subjective and presentation characteristics. Participants were then asked to complete a Dissociative Experiences Scale (DES) questionnaire to examine i f susceptibility to created memories was related to a general dissociative cognitive pattern. Results indicated that twenty (26%) o f participants created complete memories for the false emotional events (seven animal attacks, five instances o f getting seriously hurt by another chi ld , four serious outdoor accidents, three episodes o f getting lost, and one medical procedure). Furthermore, 29.9% o f participants reported some false information pertaining to the false event ("partial" memories), for a total o f 55.9% o f participants recalling information relating to the false event. T h e remaining participants (44.2%) reported no information pertaining to the false event. There were several interesting differences among the three memory types, including stress ratings, v ividness/ clarity ratings, confidence ratings, coherence, number o f details, repeated details, and memory failures. F o r example, when relating a created memory, participants were less confident and the memories were less v i v i d and detailed compared to the other memory types, but similar i n sensory components and relevancy. O n the other hand, participants were highly confident i n their fabricated memories w h i c h were rated as highly stressful and v iv id , and the memories were detailed. The Architecture of Childhood Memories iv However , w h e n relating a fabricated memory, participants repeated details and were less wi l l ing to admit lacking memory, relative to real memories. Other findings are reported o n the origin o f the created memories, age factors, memory perspective, reasons provided for first forgetting the false event, and post-interview confidence i n the created memories. O n the D E S , participants w h o had come up w i t h a partial or a created memory scored, on average, about twice as high as those participants w h o had recalled no false information, indicating that susceptibility to memory distortion may be related to a general dissociative pattern. This was the largest scale created memory study to date and the first to look at a variety o f emotional ch i ldhood events and the content o f the memories. Implications o f the findings for the false memory debate and memory assessment i n forensic contexts are discussed. The Architecture of Childhood Memories TABLE OF CONTENTS A b s t r a c t VI Table o f Contents y L i s t o f Tables Vll List o f Figures ^jjj Acknowledgments l * I N T R O D U C T I O N A K i l l e r T o m a t o attack 1 Chapter O n e The Classical Spatial V i e w o f M e m o r y 6 A n c i e n t views on memory 9 Spatial views i n later philosophy and psychology 11 Chapter T w o Trac ing the Contemporary V i e w o f M e m o r y as Architect 17 Reconstructive views before 1967 19 M e m o r y i n the Renaissance and Enlightenment periods 19 Conceptions o f memory distortion i n the 19th and early 20th centuries 24 Reconstructive views o f memory after 1967 Remembering Remembering: Neisser's reconstructive v iew o f memory 36 Cont inu ing evidence for memory reconstruction f rom cognit ive/ forensic psychology 37 A n existential reconstructivist perspective 45 Chapter Three Notable Returns to the Storehouse Travels back to the storehouse 62 " F l a s h b u l b " memories 62 T h e Repressed Memory : A prevalent modern storehouse not ion 66 Eidet ic memories 78 The strange case o f S. 79 Chapter F o u r The Architecture o f E m o t i o n a l Memories Basic factors affecting memory quality 83 T h e reconstruction o f emotional events by adults 96 The reconstruction o f emotional events by children 101 T h e reconstruction o f traumatic events 112 The Architecture of Childhood Memories yi Chapter F ive T h e Architecture o f Created Memories 125 T h e recent discovery o f Created/ Implanted memories 126 Created memories i n chi ldhood 127 Created memories i n adulthood 130 Chapter Six T h e Architecture o f Lies 151 T h e phenomenon o f lying and its forensic manifestations 152 Detect ing deceit: A n o ld problem 156 Assessing fabricated memories by their content 159 Chapter Seven T o w a r d a Resolut ion o f the Debate: The N e e d to Produce and Examine Created Memories 167 Chapter E ight T h e Present Study Examines Real, Created, and Fabricated C h i l d h o o d Memories 191 Introduction to the present study 192 Methodologica l strategy 193 Est imat ing the base rate o f implanted memories 196 E x p l o r i n g possible differences i n the content o f the memories 197 Rationale for event categories and age range 205 Operationally defining "created memories" 207 Establishing whether the emotional events occurred 208 M e t h o d 210 Results 222 Discuss ion 251 Conc lus ion 268 References 270 Appendices A C h i l d h o o d memory parental consent form and questionnaire 307 B Interviewing scripts 312 C Consent f o r m 319 D Debrief ing form 320 E E t h i c a l approval 322 Endnotes 323 The Architecture of Childhood Memories LIST OF TABLES Table 1. Spatial Conceptions o f Memory : Ma jor His tor ica l Figures and their Contributions 14 Table 2. Architectural Conceptions o f Memory : Ma jor His tor ica l Figures and their Contributions 42 Table 3. Studies o f Created/ Implanted Memories i n Adul t s 139 Table 4. Frequency o f Implanted Events and Implantation Rates 225 Table 5. T w o Examples o f Partial Memories 226 Table 6. T w o Examples o f Created Memories 227 Table 7. Inter-Rater Reliabilities for the Content Criteria 230 Table 8. Summary o f the Analysis Compar ing Real, Created, and Fabricated Memories 239 Table 9. Summary o f the Analysis Compar ing Real and Fabricated Memories Across the Sample 241 Table 10. Reasons Participants Gave for N o t Initially Recall ing the False E v e n t 246 The Architecture of Childhood Memories vYi'i LIST OF FIGURES Figure 1. Propor t ion o f Participants w h o Recalled a Created M e m o r y , Partial M e m o r y , or N o M e m o r y 224 Figure 2. Stress Ratings for the Real, Fabricated, and Created Memories 234 Figure 3. Vividness Ratings for the Real, Fabricated, and Created Memories 235 Figure 4. Confidence Ratings for the Real, Fabricated, and Created Memories 236 Figure 5. Coherence Ratings for the Real, Fabricated, and Created Memories 237 Figure 6. N u m b e r o f Details i n the Real, Fabricated, and Created Memories 238 Figure 7. N u m b e r o f Repeated Details i n the Real, Fabricated, and Created Memories 241 Figure 8. N u m b e r o f M e m o r y Failures i n the Real, Fabricated, and Created Memories 242 Figure 9. D o m i n a n t M e m o r y Perspective i n the Real, Fabricated, and Created Memories 244 Figure 10. Dissociative Experiences Scale (DES) Scores i n the N o M e m o r y , Partial M e m o r y , and Created M e m o r y Groups 249 Figure 11. D E S Scores i n the N o M e m o r y Dis tor t ion and Poo led M e m o r y Di s tor t ion G r o u p s 250 The Architecture of Childhood Memories ix ACKNOWLEDGMENTS I w o u l d like to thank all members o f my P h . D . Committee , i n particular Drs . Robert Hare and D a r r i n L e h m a n for great feedback at earlier stages o f the dissertation. I found an exceptional group o f researchers, colleagues, and friends i n D a r r i n Lehman's Advanced Social Psychological Research 408 class (Andrea, Bobby , Danae, E v a , G w y n , Harpreet, K a m i , Manendra, and Rasool), w h o contributed a tremendous amount o f their time, energy, and ideas as interviewers and methodological consultants throughout the 1997-98 school year. Thanks to Diane , A n d r i i , and V i c k y for the t ime-consuming but excellent coding job. Danae Drab contributed an exceptional amount o f her time and support, and helped create an organizational tour de force during data collection. I also thank D a n a Thordarson for her help w i t h statistical consulting. I thank D r . J o h n Yui l le for his excellent supervision throughout my doctoral studies and, especially, my trout fishing studies. Thanks to Correct ional Services Canada (CSC) for graciously al lowing me research time to complete this study during my time as a Psychological Associate. Apprec ia t ion is extended to the Social Sciences and Humanit ies Research C o u n c i l o f Canada (SSHRC) and the A m e r i c a n Psychology and L a w Society ( A P L S : A P A D i v . 41) for financially supporting this dissertation. Finally, I w o u l d like to thank Ange la B i r t (aka Bertrand) for her love, support, and her o w n endless supply o f created memories. The Architecture of Childhood Memories 1 A n Architectural M i n d : T h e Nature o f Real , Created, and Fabricated Memories for E m o t i o n a l C h i l d h o o d Events I held a jewel i n my fingers A n d went to sleep. T h e day was warm, and winds were prosy I said: ' T w i l l keep.' I woke and chid my honest fingers,-T h e gem was gone; A n d n o w an amethyst remembrance Is all I own. -Emi ly D i c k i n s o n (1959) A K i l l e r T o m a t o Attack O n a recent return to my hometown, my sense o f reality was undermined during a discussion o f a personally traumatic event w i t h a close friend. In the course o f the chat, a bizarre and highly memorable incident w h i c h had occurred twelve years before came up. The way I was telling it, the friend and I had been strolling d o w n a street i n a litde N o v a Scotian town when a shot rang out. I vividly remembered my friend a few steps ahead o f me and w h o , o n hearing the shot, immediately turned to see i f I was okay. Seeing my chest, her face paled w h i c h prompted me to look d o w n at my T-shirt, a sight I w i l l never forget. O v e r the emaciated bodies o f the Jagger and Richards, meandering tributaries were f lowing f rom three great pools of crimson. Real iz ing that I had been struck by a shotgun blast, I keeled over i n shock while she shouted at the top o f her lungs for help. A s she frantically tried to beckon a car, she noticed something o n the ground near the crime scene and then eyed my wounds more closely. A s I reflected back over my life, p ining its brevity, and preparing to meet my maker, her tears f lowed not to existential cries but rather to the tune o f hearty laughter. She held up the remnants o f the The Architecture of Childhood Memories 2 most unusual weapon - a packet labeled " H e i n z K e t c h u p . " O n this sweltering summer day, she had stepped o n the expanded packet w i t h a heel and the contents had sprayed back on yours truly. A humbl ing experience indeed. Af ter recounting this version to my patient friend, she declared that I had the whole story backwards. Rather, it was she w h o had been the recipient o f the ketchup attack and I the heavy-heeled perpetrator. A s I had told this traumatic tale o n numerous occasions, I could claim the accurate version w i t h a high degree o f certainty. However , a second friend intervened at this point to in formed us that I had related her story to h i m shortly after the event happened twelve years before. I trust my inaccuracy was not the result o f vacillating sanity or early-onset Alzheimer 's . Instead, it demonstrates the reconstruction o f past life events. In a formal survey w i t h my friends as participants, it appears that everyone could recount an occurrence replete w i t h wrongful retrospection, especially chi ldhood events. In a study by K a r r (1990), 87% o f participants could describe at least one instance i n w h i c h memory for a personal incident conflicted w i t h someone else's. M o r e recently, intriguing research by Loftus and colleagues has demonstrated that some people can be led to recall ch i ldhood events w h i c h never actually occurred (see Loftus , 1997a, 1997b), although m u c h controversy exists about what types o f events (see K n a p p & Vandcreek, 1997; P e z d e k & Roe , 1997; Read & Lindsay, 1997). H o w can people recall life's events so differently f rom their objective reality and how do we revise our personal histories? Is memory hopelessly fallible or, as one o f Gogol ' s characters tells us, "just like pour ing water through a sieve" (Gogol , 1832/ 1957, p. 63). A n d is it possible to k n o w whether ch i ldhood memories are founded i n reality? In forensic and clinical settings, professionals and other decision-makers c o m m o n l y evaluate the validity o f testimony for emotional chi ldhood events. There are three possibilities The Architecture of Childhood Memories 3 to consider w h e n someone relates a memory: (1) that the event actually occurred; (2) that the event d id not occur but the person thinks that it d id (a "created" memory) ; and (3) that the event d id not occur but the person knows it is false (a fabricated memory). That is, the reported memory could be based o n a real event, an imagined event, or a made-up event. In particular, it is the second possibility that has been the subject o f an ongoing heated debate i n psychology and the courtroom (e.g., Pope, 1996). O n e aspect o f the debate addresses whether it is possible to "create" or implant false chi ldhood memories i n a person (for example, by therapists or police), and i f so, the nature and incidence o f such created memories (e.g., Lindsay & Read, 1994; Loftus , 1997b). Further, are there differences i n h o w a person recalls and reports these three types o f memories? A l t h o u g h this area o f psychology is still i n its infancy, i n a short period researchers f rom many areas o f psychology, including forensic, cognitive, and neuropsychology have been applying their expertise to investigating memory distortion, including created or implanted memories (e.g., Loftus, 1997a, 1997c; Schacter, 1995a), confirmed i n the several recent texts devoted to the topic (e.g., Read & Lindsay, 1997; Schacter, 1995a, 1996; T h o m p s o n , Herrman , Read, Bruce, Payne, & Togl ia , 1998). A s wel l , A m e r i c a n Psychologist has been the forum for an ongoing written debate beginning w i t h the critique by leading A m e r i c a n ethicist K e n n e t h Pope (1996) o n the research and lack thereof o n w h i c h the so-called false memory syndrome concept is based. Pope argued that the debate has been marred by politics and polemics, and called for a cautious, dispassionate empirical approach to examining the phenomenon o f false memories and the use o f various suggestive therapeutic practices. O n e o f Pope's most important points is that the strong conclusions about false memories made by many psychologists (such as El izabeth Loftus) are based o n a very l imited number o f studies each w i t h shortcomings. A l t h o u g h a great deal has been published and stated o n bo th created memories and recovered memories, little has been empirically based (e.g., G o l d , The Architecture of Childhood Memories 4 1997). A s Pope (1996) pointed out, "each scientific claim should prevail or fall o n its research validation and log i c " (p. 971). Clearly, given the stakes for psychologists, psychologists' clients, complainants o f and those o f accused o f historical chi ld sexual abuse (CSA) , m u c h more research is needed to answer the complex issues at the heart o f the debate. T h e present study represents one o f the first direct explorations o f the phenomenon o f created memories for emotional ch i ldhood events, and the first to investigate lying and created memories simultaneously. A l t h o u g h memories for false chi ldhood events have l o n g been k n o w n as "false" memories, the term was used infrequendy i n this paper. It may be that these memories are as subjectively real to the person as real memories. Whatever the objective reality o f a particular event, i f the person remembers it then it is not truly a false memory. Secondly, as I w i l l argue throughout this dissertation, reconstruction and distortion may be basic to memory for events and all memories may thus be "false." Throughout this paper, an attempt w i l l be made to consistently refer to memories w h i c h are not based on an objective external event as "created" memories to communicate that they have been constructed f rom something other than reality (e.g., Porter & Marxsen, 1998), or, alternatively, " implanted memories " w h e n the memories were specifically created by an outside person (e.g., an experimenter). F o r simplicity, memories based o n objective reality w i l l be called real memories and lied-about memories w i l l be referred to as fabricated memories. A n init ial agenda in this dissertation was to investigate how people recall events o f high emotional significance. A discourse is provided o n the reconstructive nature o f autobiographical memory, a relatively recent v iew calling into question the millennia-old storehouse metaphor. Research and theory o n autobiographical memory (with a focus on ch i ldhood events at the high end o f the emotional significance spectrum) is reviewed. It is concluded that memories for meaningful life events are reconstructed rather than reinstated, that reconstruction o f such events is The Architecture of Childhood Memories 5 generally a rational, reliable, and powerful process. However , distortion is also characteristic o f memory, is generally adaptive, but can sometimes culminate i n a memory image being constructed for an entire event wi th no objective external reality (i.e., a created memory). Further, the reconstructive aspects o f memories are generally created unintentionally but one form o f memory is created intentionally - the lie. The idea o f whether the three memory types can be discriminated is explored. A d o p t i n g a broad framework designated existential reconstructivism, the many existing ideas o n h o w people recall meaningful life events (notably flashbulb memories, remarkable memories, repressed memories), and false events, are discussed. The evidence indicates that autobiographical memory is primarily a reconstructive process. People reconstruct and sometimes deconstruct events, and events are likely never recalled exactly as they were originally experienced. It appears that gradations o f accuracy result f rom many factors but foremost may be depth and frequency o f reflection o n the event, privately (e.g., tiiinking or ckearning about the event) or socially (telling others about the event). D e p t h and frequency o f reflection are generally dependent o n their perceived adaptive value or pertinence to personal and others' well-being {existentialroles). It is possible that memories grounded i n reality w i l l be qualitatively different f rom those grounded i n imagination, an eighteenth century proposal o f Bri t i sh philosopher, D a v i d H u m e , and reiterated by twentieth century G e r m a n psychologist U d o Undeutsch. T h e essential matter o f this dissertation, then, was to provide the first i n depth exploration o f the nature o f real, created, and fabricated emotional ch i ldhood memories. The Architecture of Childhood Memories CHAPTER ONE The Classical Spatial View of Memory The Architecture of Childhood Memories 7 M e m o r y has always been a foremost concern o f philosophers, psychologists, and laypersons. R i c h , enlightening discussions have been penned o n this most fundamental aspect o f our mental life. Dostoyevsky argued that people cannot live without memories, especially the precious memories o f chi ldhood. Despite its importance and the extent o f the attention it has attracted, most psychological research on memory processes i n the past century has been model led after Ebbinghaus ' init ial laboratory experiments, tighdy controlled and invo lv ing basic stimuli (e.g., K o r i a t & Go ld smi th , 1996). This restricted approach has been crit icized (e.g., Neisser, 1978, 1982), w i t h calls for more ecologically val id "real w o r l d " memory studies (cf. Banaji & Crowder , 1989). Neisser (1982) has called for the study o f topics having greater applicability, not ing that "we have almost no systematic knowledge about memory as it occurs i n the course o f everyday l i f e " (p. xii) and, i n 1981, that there has been little sense o f progress since the early experiments, w i t h most research focussing on methodological issues. Thi s concern has been reiterated more fully i n a text edited by Gruneberg and Morr i s (1992). T h e response has been the generation o f a large sub-area o f memory research, the study o f "autobiographical m e m o r y " (e.g., see Neisser, 1982), including more realistic studies addressing important memory distortion issues wi th in forensic psychology. This recognition recently led to the first meeting o f the Society for A p p l i e d Research in M e m o r y and Cogni t ion ( S A R M A C ) , an international conference devoted to applied memory issues, held i n Vancouver , B C , in 1995. M o r e specifically, the effect o f emotion on autobiographical memory has been a major focus i n psychology i n the past few years (e.g., Christianson, 1992). O u t o f this changing Zeitgeist, the phenomenon o f created memories was first explored (see Read & Lindsay, 1997), w i t h considerable implications for our understanding o f human psychology generally. W i t h questions arising i n the false memory debate in m i n d , the present paper seeks to explore and shed light on the business o f h o w the m i n d constructs memories for emotional The Architecture of Childhood Memories 8 chi ldhood events. However , the psychological literature is replete w i t h ideas o f h o w we recall emotional events. Evaluating all o f them i n a single paper w o u l d be exceedingly difficult, so, here, the analysis is l imited to: (1) " forgot ten" classical ideas about the qualities o f memory; (2) ideas about memories for events o f high emotional significance and; (3) those concepts w h i c h have received the greatest theoretical and empirical attention. Rather than constricting the discussion o f memory to the successes, failures, and trends i n modern psychology, the conceptions o f phi losophical thinkers f rom ancient times are explored and the ideas and research o f early psychology are revisited, as each has pertinence to understanding memory construction. A n awareness o f the history o f predominant thought o n the nature o f memory is important, for three main reasons (not to mention the inherent interest i n the historical stroll), as outlined by H e r r m a n n and Chaff in (1988). First, an understanding o f the classic literature on the qualities o f memory can facilitate modern scholarship i n psychology by prov id ing a better appreciation o f the current thinking i n the study o f memory. Secondly, it can provide insights into memory. Current views o f memory owe m u c h to discerning observations (and the equally discerning /^observat ions) through the course o f human history. Finally, it circumvents a c o m m o n problem - that o f " forgett ing" already-discovered things about memory as noted by Neisser (1982), those w h o ignore memory's history are condemned to repeat it. Forensic psychology seems to have disregarded earlier observations o n memory w h i c h are pertinent to the task at hand - assessing the validity o f memories for emotional events. A final point is that, as one Russian novelist (Salaman, 1970) has observed, philosophers and writers k n o w things about memories and their creative aspects o f w h i c h psychologists may not, point ing out the importance o f considering historical philosophical and literary views o n memory. P r i o r to this century, the prevailing philosophical v iew was that memory was a spatial entity i n w h i c h events were kept intact somewhere i n the m i n d , akin to a storehouse (e.g., The Architecture of Childhood Memories 9 H e r r m a n n & Chaf f in , 1988). F o r thousands o f years, writers conceptualized memory as a virtual reproduction o f original sensory experience, or at least faded reproduction. Thi s conception o f mind as a mental space, i n w h i c h memories are stored and retrieved by search, has served as a general and powerful explanation o f memory (e.g., Roediger, 1980). T h e origins o f spatial views are explored next and later considered i n h o w the m i n d can create memories for things w h i c h never occurred. A n c i e n t V iews on M e m o r y The ancient Greeks devoted m u c h spiritual and philosophical meditation to the subject o f remembering. O n e o f the twelve Titans, the great primeval deities o f nature, was Mnemosyne , the goddess o f memory, necessary for all mental activities (Hamil ton, 1969; Lines , 1973). T h e pervasive v iew was that memories were stored intact i n the m i n d and one's ability to retrieve them was a measure o f intellect. Centuries before the next major memory theorist (Augustine), memory had been viewed as a cavern o f wax by Socrates, Plato, and Ar i s tode (see Table 1). Socrates held the belief that there is a wax tablet o n the m i n d " w h i c h is larger i n one person, smaller i n another, o f pure wax i n one, more muddy i n another, harder i n some, softer i n others, and i n some just r ight" (Socrates, 369 B C / 1972, p. 186). D r a w i n g o n the ideas o f his two predecessors, Ar i s tode advanced the theory that memory and recollection are distinctive (Aristode, c. 335 B C / 1988). V i e w i n g the formation o f memory, or sensory stimulation, as resulting f rom the mark o f perception, he commented: "the act o f perception stamps i n , as it were, a sort o f impression o f the percept (on the soul), just as persons w h o make an impression w i t h a seal" (p. 65). People wi th poor memories, such as the young and o ld , had defective receiving surfaces o n their souls, as i f the seal were stamped o n a water surface or frayed stucco (for young and o ld , respectively). M e m o r y was contingent on the quality o f perception and then The Architecture of Childhood Memories 10 the ability to retrieve the perceived material. This v iew o f his three intellectual forefathers failed to wax i n the m i n d o f Plotinus (205-270 A D ) , the "last great philosopher o f antiquity" (Russell, 1945) and one o f very few i n the thousand years after them to object. T h e Enneads (c. 244 A D / 1962) had the dissenting chapter title, " M e m o r y is not passive, but is an active power o f the m i n d , " i n w h i c h Plotinus argued that sensory perceptions are nothing like imprints and, secondly, that memory is a rational force o f the soul, "the reasoning power o f the universe." Despite its ethereal quality, his discussion marks an important early transition f rom memory as spatial to memory as constructive. Re-asserting the spatial memory idea, Augustine (354-430 A D ) introduced the wel l -k n o w n and influential "storehouse" metaphor. In his defining work, Augustine: Confessions and E n c h i r i d i o n (in H e r r m a n n & Chaff in , 1988), while purging the v i v i d memories o f his youthful transgressions, Augustine argued that all emotional events are encoded and permanently stored i n original form, an idea w h i c h w o u l d continue to permeate phi losophical thought for centuries. Augustine wrote o f the "vast cave o f memory w i t h its numerous and mysterious recesses" w h i c h receives events and "stores them up, to be recalled and brought forth when required" (p. 112). In the tradition o f Aristotelian principles, memory failures had their origins i n retrieval difficulties. A c c o r d i n g to these early wax and storehouse views, the qualities o f reinstated memories i n competent rememberers should be similar to (or at least an inverted reflection of) the events as originally perceived. F r o m Augustine, stressful or emotional life events could be retrieved and literally communicated to another person. The nature o f an event could thus be inferred f rom direct memory experience or f rom listening to a person's verbal description, w i t h appropriate attention to the mental characteristics o f the rememberer. Throughout this period, no writers are k n o w n to have considered or observed entire created memories for false events. The Architecture of Childhood Memories 11 Spatial V i e w s i n Later Phi losophy and Psychology T h e D a r k Ages were a barren time for w o r k o n memory and most everything else. F o l l o w i n g the treatises o f Augustine, writings o n memory disappeared unt i l a trickle was created by Thomas Aquinas , w h i c h quickly grew into a river o f thought o n the matter i n the Renaissance. F r o m the sixteenth to the eighteenth centuries, two loosely integrated theoretical schools o f memory existed. W h i l e the first held tenaciously onto the classical Aristotel ian and August inian views, the other advocated a liberal v iew i n line w i t h what was becoming the sovereign emphasis o n reason. Aristode's passive wax not ion o f memory was adopted by many during this period, notably J o h n L o c k e and D a v i d H u m e . L i k e Augustine, L o c k e believed there to be nothing i n the intellect except what had been originally experienced through the senses (Durant, 1961), a statement to w h i c h Leibni tz , i n the Enlightenment period, appended: nothing except the intellect itself. Associationists i n the nineteenth and twentieth centuries believed that a memory trace remains permanendy intact after it becomes part o f long-term memory. A c c o r d i n g to W i l l i a m James, an impression could be sufficiendy emotionally arousing to almost "leave a scar u p o n the cerebral tissues" (1890/ 1950, p. 670). The late nineteenth century witnessed the advent o f psychoanalysis, a school w h i c h advocated the spatially-oriented idea that certain ch i ldhood events become repressed or buried i n the unconscious. Repression was conceived by Freud as a back room, or "ante-room", guarded by a door-keeper, (discussed later). In an essay o n French philosopher H e n r i Bergson and Bergson's views o n memory, Bertrand Russell (1945) encapsulated spatial views: It is suggested that i n some way everything that has happened to us is remembered, but as a rule, only what is useful comes into consciousness. Apparent failures o f memory, it The Architecture of Childhood Memories 12 is argued, are not really failures o f the mental part o f memory, but o f the motor mechanism for bringing memory into action (p. 797). In the 1950s, famed neurologist, Wi lder Penfield stimulated the axons o f the scientific community to strongly embrace this storehouse not ion (see Table 1). H e reported that long-buried memories could be elicited by stimulation o f localized brain sites, particularly i n the temporal lobes (Penfield, 1952). Af ter asserting that memory was a reproduction or reinstatement o f the original sensory experience, he was prompted to state, " I t has fallen to my lot, during explorations o f the cortex, to demonstrate a mechanism i n the human brain w h i c h preserves the record o f the stream o f thought" (Penfield, 1954, p. 47). In the same year as Penfield's inaugural publication on memory permanence, Federn (1952) highlighted the psychoanalytic v iew o f memory - that chi ldhood experiences remained preserved i n a latent state unti l re-experienced (a v iew repeated by Berne, 1961). In an influential paper, A t k i n s o n and Shiffrin (1968) argued that long-term memory is a permanent storage system and a variable, or labile, retrieval system. However , i n the next thirty years, a wealth o f research suggested that reconstructive forces influence memory and that there is m u c h more to memory quality than retrieval capacity. Nonetheless, during that time, a number o f academics and mental health professionals continued to wander around the dusky memory storehouse. Wickelgren (1977) weighed the evidence for the reconstruction view and concluded, i n an Aristotel ian argument: There is no evidence against qualitative change i n memory, but there is also no evidence for it, and there is still no reason to doubt the parsimonious hypothesis that only quantitative changes occur i n memory storage i n the absence o f new learning or rehearsal.. .there is no reason to doubt the parsimonious hypothesis that changes i n trace strength are strictly negative...So far as we know, time does not "rewrite the l ines" i n our memories, it only makes what is written o n some lines fade (p. 392). The Architecture of Childhood Memories 13 F r o m a clinician's point o f view, L e v i n (1985) stated even more recendy that all early experiences are recorded i n ego states called the Natura l C h i l d , " o n f i lm and o n file i n each o f us" (p. 95). Psychology Today further upheld the storehouse v iew i n a 1988 advertisement: "Science has proven the accumulated experience o f a lifetime is perfectly preserved i n your m i n d " (in Myers , 1990, p. 112). L a m a l (1979) reported that 85% o f a sample o f A m e r i c a n university students (who allegedly subscribed to Psychology Today) agreed that memory was akin to a storage chest. This discussion was intended to survey the dominant spatial v iew o f memory through the history o f Western philosophy and psychology, summarized i n Table 1. F r o m ancient times memory was v iewed spatially, w i t h the exception o f iconoclasts like Plotinus. It was thought that the original perceptions o f events were stored i n space wi th in the m i n d or soul. T h e Greeks embraced the "passive w a x " not ion, evolving into the cavern/ storehouse metaphor o f Augustine and subsequent thinkers (e.g., Roediger, 1980). Associationists thought that memory traces remained intact after becoming part o f long-term memory. Late i n the nineteenth century, a central tenet o f psychoanalysis was that certain chi ldhood memories can remain intact i n the unconscious indefinitely. Penfield's experiments served to reify for the scientific community the spatial metaphor by correlating the stimulation o f particular brain sites i n epileptics w i t h the elicitation o f " long-bur ied" autobiographical memories. W h y was the spatial v iew that memory resides i n the m i n d or brain so enduring throughout history? A simplified answer might be that it makes intuitive sense, particularly i n the (sometimes dogmatic) context o f the long-standing ideas on overarching spiritual and metaphysical questions. The soul, m i n d , and even the afterworld were seen i n many cultures as literal places. T h e nether regions o f the mysterious soul were where the essence o f man - good The Architecture of Childhood Memories 14 and evi l - resided, die mysterious afterlife was where souls travelled, and the mysterious m i n d was where thoughts and memories were stored. In addition to this general spatial outlook o n life, the subjective experience o f spontaneously remembering an event v iv idly or hearing a v i v i d description o f a memory led thinkers to conceive o f memories as entities w h i c h are stored intact, a v iew w h i c h has sometimes re-surfaced i n modern psychology. F r o m this spatial perspective, what were the qualities o f our autobiographical memories? Clearly, i f memory and perception were essentially the same (memory was a copy o f what was perceived) then the question was answered simply by Hstening to the person's description o f the event (assuming the person was a competent rememberer). Table 1 Spatial Conceptions o f Memory : Ma jor Flistorical Figures and Their Contr ibut ions His tor ica l figure/ perspective Contr ibut ion Socrates (c. 369 B C ) Aristot le (c. 3 3 5 / 1908 B C ) Augustine (c. 400 A D ) J o h n L o c k e , D a v i d H u m e Associationists (e.g., James) Perception occurs i n a wax tablet o n the m i n d . M e m o r y represents a reading o f the tablet. Perception is an impression made w i t h a seal. H e also formulated the first associationistic principles. H e first discussed memory as a "storehouse." M e m o r y is reinstated perception. H u m e discussed differences between memory and imagination. The memory trace remains intact after entering L T M . James also discussed the effect o f emot ion o n memory (it could leave H e n r i Bergson (late 19th and early 20th century writings) Psychoanalysis (e.g., Freud, 1922) Eidet ic memories (e.g., K l i i v e r , 1926) Wi lder Penfield (1952, 1954) B r o w n & K u l i k (1977) The Architecture of Childhood Memories 15 a "scar on the tissues"). All events are stored as perceived. M e m o r y failures are due to retrieval failures. Discussed by Schopenhauer and adopted by Freud, repression is an explicit spatial conception o f memory. Painful ch i ldhood events are kept out o f conscious awareness by remaining intact i n the "ante-room", or unconscious unt i l they can be brought to the surface. Advocates o f this idea held that some people's memory imagery is characterized by an incredibly v i v i d , literal reproduction o f the original event. Eventually, it was shown that most eidetic memories were, i n fact, not veridical and sometimes, inaccurate. T h r o u g h experiments o n conscious epileptics' brains, he came to report that all events are stored as originally perceived i n localized brain regions and memory works like a video camera (similar to Bergson's ideas). A l t h o u g h his experiments immediately made waves i n the scientific community , his idea was later refuted. They re-popularized the storehouse not ion o f memory w i t h their classic paper on " f lashbulb" memories - v i v i d , accurate memories for culturally significant events, such as polit ical Cognit ive Science/ Psychology (1960s to present) The Architecture of Childhood Memories 16 assassinations. They asserted that the details o f the original context o f hearing o f such an event become " f r o z e n " i n a person's m i n d . Later studies (e.g., Neisser & Har sh , 1992) amassed evidence w h i c h refuted the idea. M e m o r y metaphors included libraries, computers, fi l ing systems, etc. (e.g., the moda l mode l o f memory). The Architecture of Childhood Memories 17 CHAPTER TWO Tracing the Contemporary View of Memory as Architect The Architecture of Childhood Memories 18 T i m e and memory are true artists; they remould reality nearer to the heart's desire (Dewey, 1920). A s mentioned, a belief i n the potential storage and recoverability o f ch i ldhood memories continues i n some modern perspectives, exemplified by its central role i n contemporary psychoanalysis, and an idea attacked by the false memory faction i n the debate (Brown, Scheflin, & H a m m o n d , 1998). The storehouse idea has sometimes been given credence by modern researchers o f a cognitive bent (Table 1). F o r example, psychologists have c o m m o n l y used computer analogies, have considered the not ion that memory involves the stacking o f things i n layers so that the "bur ied" items are more difficult to access (e.g., Bekerian & Bowers , 1983), and have argued the idea o f " f lashbulb" memories. B u t it turns out that time - or more specifically forces contingent o n time - does "re-write the l ines" i n our memories. W i t h his quote above, J o h n D e w e y observed that memory is not a passive, objective receptacle o f perceptions i n accordance w i t h the classical posit ion but rather is active and subjective, expanding o n Plot inus ' view. It should also be noted that although Aristotle embraced the passive wax idea o f memory, he had also viewed recollection as an inferential process dependent on currently-held knowledge, a k i n d o f investigation guided by the rules o f association. Aristotle felt that a person searches for the original image using a logical search (Aristotle, trans. 1908); that is, he viewed the recollection o f an event as contingent on the principle o f rationality (Brennan, 1941). Further, although it was earlier pointed out that created memories had not been considered explicitly unti l modern times, both Socrates, as related by Plato, and Plato himself (in T h e Sophist) had addressed at length the problem offalse beliefs (although not false memories) i n their writings o n epistemology (see Matthews, 1972). Spatial metaphors o f memory have been progressively supplanted by a reconstructive or architectural v iew, n o w influencing m u c h w o r k o n memory i n psychology and phi losophy (e.g., The Architecture of Childhood Memories 19 K o r i a t & G o l d s m i t h , 1996; Payne, Neuschatz, Lampinen , & L y n n , 1997). A l t h o u g h modern views o n event reconstruction are considerably different f rom the classical v iew, philosophers have long-recognized that memory is subject to distortion, despite adhering to variations o f the storehouse concept. A t least since the ancient Greeks reconstructive forces were sometimes (but uncommonly) observed. Twentieth century psychology has witnessed a gradual shift i n our understanding o f h o w people recall life events (see K o r i a t & G o l d s m i t h , 1996; Neisser, 1967). M a n y memory theorists have begun to move past traditional spatial views. Reconstruction is n o w viewed by some as a major factor i n the growth o f memories and the original perception only the seed. I. Reconstructive Views Before 1967 M e m o r y i n the Renaissance and Enl ightenment Periods Insights o n memory, like scientific discoveries i n general, have never occurred i n a vacuum. D u r i n g the Renaissance period i n Europe , many thinkers were re-considering and challenging long-held classical views o f human nature, i n the spirit o f the revolutionary social and intellectual climate. Ha l lowed Hippocrat ic and Galenic conceptions o f the human body were challenged by Vesalius and Harvey. Aristotel ian views o n phi losophical investigation were challenged by Bacon , Gal i leo , and Descartes. A n d the uniquely human capacity o f reason was at the essence o f scientific explanations and endeavors (Voltaire, 1766/ 1965). W i t h the Reformation, the ecclesiastic, conventional dogma w h i c h had clouded scientific consciousness for so long began to dissipate. T h e unmasking o f nature by Copernicus, Descartes, and other thinkers, then, led people to question the unquestionable and obvious. T h e Aristotel ian " t r u t h " o f memory was no exception. Great artistic achievements o f this time no doubt too contributed to the changing perspective. Renaissance artistic works represented products o f the human The Architecture of Childhood Memories 20 m i n d , an alliance o f the artist's personal experiences w i t h h i s / her reasoning and creativity. Maybe memory operated under a similar principle. It was i n this atmosphere that the seed o f a major evolution o f memory theory formed, developing into the second theoretical "tree." D u r i n g and after the A g e o f Reason period o f Europe , philosophers re-discovered and expanded the insights o f Plot inus and disputed stricdy spatial views. In this period, memory as a passive receptacle or reflection o f sensory inputs instead came by some to be seen as an active, creative force governed by reason, as Plotinus had asserted. Memories were no longer assumed to necessarily reside " somewhere" i n the m i n d or soul. Still , the classical spatial v iew was not rejected overnight or by most thinkers. Rather, it was steadfasdy maintained by major thinkers such as H u m e , and Descartes, the latter w h o , i n his writings on dualism, described most mental functions such as c o m m o n sense, imagination, and memory using the wax metaphor (Wilson, 1969). M e m o r y as constructive reasoning: Radical proposal o f Le ibni tz and K a n t A key doctrine o f the first modern pre-eminent storehouse antagonists was that memory is contingent o n reason and has active complexity. T w o G e r m a n philosophers - Le ibni tz and K a n t - put forth this idea that memory was active rather than passive, one still o n the intellectual fringe during the Enlightenment. B y formulating this concept, Le ibni tz and K a n t thereafter spurred a re-di inking o f memory i n w h i c h the storehouse assumption was challenged. Despite v iewing memory as an act o f the soul, Le ibni tz (1704/ 1908) observed the importance o f active reasoning and interpretation o n memory for our experiences. H e apparendy recognized the influence o f social factors on memory, evidenced i n his statement, " . . .1 might remember more distant things, the testimony o f others might f i l l the gap o f m y remembrance" (p. 231). A c c o r d i n g to Le ibni tz , problems o f mora l identity could arise i n the face o f memory The Architecture of Childhood Memories 21 m f k m i t y and provided an early acknowledgment o f created memories: " I t is true that i f others conspired to deceive me (as I might even be deceived by myself, by some v i s ion , dream or illness, believing that what I dreamed had happened to me), the appearance w o u l d be false" (p. 231). (Interestingly, Le ibni tz [1701/1908] also gave an early discussion o f psychological processes occurr ing outside o f consciousness). N o t only did Leibni tz challenge that memory capacity is related solely to retrieval ability, he also introduced some possible distorting influences o n memory. These ideas o f Leibni tz were raised to great heights by Immanuel K a n t i n his Crit ique o f Pure Reason (1781/ 1966). K a n t proposed that our experiences in fo rm us about existence, but only through separate sensations and occurrences. In trying to establish truths about ourselves and the nature o f things i n general, humans employ reasoning processes to interpret and organize sensory experiences. Rejecting the premise o f Aristot le , K a n t d id not see the human m i n d as being like passive wax upon w h i c h sensory experiences leave their indelible marks, nor d id he espouse Augustine's explicit storehouse metaphor. H i s "last great thesis" was that the m i n d was an active entity shaping and coordinating sensations into thoughts, ideas, and memories (Durant, 1961). K a n t held a hierarchical v iew o f the human mind : sensation is unorganized stimulation (Empfindung), perception is organized sensations (Anschauungen), knowledge/ memory is organized and constructed perception (Verstand), science is organized knowledge, and w i s d o m is organized life. L i k e the Gestaltist theorists to come, K a n t saw the essence o f each rung as more than the sum o f its elements. The intellect was an active, creative, reasoning force. T h e faculty o f memory was no exception; he viewed it as creative and by no means just reinstated perception. K a n t also advanced the idea that objects i n a person's environment have no objective reality - they only become real u p o n being k n o w n by the person. The "product ive imaginat ion" functions to "change the mani fo ld o f intuit ion into an image; it The Architecture of Childhood Memories 22 must therefore first receive the impressions into its activity, w h i c h I call to apprehend" (Kant, 1781/1966, p. 112). B y intuit ion K a n t appeared to mean to discover the past through reason facilitated by current knowledge. W i t h his great publication, K a n t turned the w o r l d o f phi losophy upside d o w n but, most importandy for our present purposes, also highlighted the constructive nature o f memory for a wider audience. F r o m the Enl ightenment literature, what could be inferred about the phenomenological characteristics o f memories? First, memories were not equivalent to perceived images, not just reinstatement, but created images resulting f rom perception and contingent o n a rational cognitive process. In this view, memory was more individualistic and less objective than impl ied i n the classical view. T h e quality o f the image would , i n addition to the original perception, reflect the reasoning capacities and experiences o f the person, independent o f the event i n question. O f course, this is not to say then that memory was independent o f perception. Perception, influenced by schemata (discussed later), was for the environment but, paradoxically, memory led to the creation o f the event, i n Kant ' s view. Leibni tz advanced that external post-event informat ion could influence the content o f the memory. Thus , the memory image for the event itself - the object o f rational reconstruction - was architectural and personal, a mental product. H u m e considers images o f memory and imagination In the same century as the emergence o f Le ibni tz ' and Kant ' s writings i n Germany, the ideas o n memory i n Bri t i sh philosophy were also evolving. D a v i d H u m e represented the threshold o f the classical spatial perspective and modern perspective on memory construction. H e advocated the wax idea o f memory but also observed the power o f the imagination to create memory-like images i n the mind . H u m e was one o f the earliest writers to explicidy contemplate The Architecture of Childhood Memories 23 the qualities o f memory and imagination, and some o f his insights have been explored by modern memory researcher Marc ia Johnson and her colleagues i n their studies o f h o w people monitor the reality o f their experiences (discussed later). Showing his reverence for Aristot le , H u m e described perception as a process i n w h i c h an impression is made u p o n the senses after w h i c h the m i n d created a copy called an " idea . " O v e r time, the idea remained i n essentially original fo rm but as faint images, faded as dust on a portrait. Impression and idea differed only i n their " l ivel iness ." In his three-volume Treatise on H u m a n Nature (1737/ 1886) (accomplished w h e n he was 26 years old), H u m e also asserted that imagined events were inferior i n their vividness and endurance to memories for real events: 'Tis evident, at first sight, that the ideas o f the memory are m u c h more lively and strong than those o f the imagination, and the former faculty paints its objects i n more distinct colours, than any w h i c h are employed by the latter. W h e n we remember any past event, the idea o f it flows i n u p o n the m i n d i n a forcible manner; whereas, i n the imagination, the perception is faint and languid, and cannot, without difficulty, be preserved by the m i n d steady and un i form for any considerable t ime" (p. 173). T h e "ideas o f memory" , i n turn, possessed "a higher degree o f v ivaci ty" and hence superiority to the " complex ideas" o f imagination. Complex imaginal ideas d id not necessarily resemble any original impressions (one can imagine a winged horse without having seen one) but the original components o f the complex idea were derived f rom experiences (Hume, 1739/ 1964, pp. 245-246). M a n y o f Hume's ideas were founded i n those o f his predecessors. H e appears to have adopted this memory decay not ion f rom Thomas Hobbes , w h o had discussed it i n Leviathan i n 1651. H i s " l ively and s trong" idea was borrowed verbatim f rom J o h n L o c k e w h o i n turn had adopted it f rom Berkeley. Berkeley had maintained that memories are more strong, lively, and distinct than ideas o f the imagination. Further, Berkeley argued that memories are The Architecture of Childhood Memories 24 characterized by a steadiness, order, and coherence not existing i n the imagination (cited i n H u m e , 1739/ 1964, p. 138) - a presage to a modern empirical f inding that real memories indeed appear to be more coherent than memories for non-experienced events (e.g., Porter & Y u i l l e , 1996). However , the dichotomies o f impressions/ ideas and m e m o r y / imagination were o f considerable significance i n Hume' s philosophical system. L i k e Berkeley, H u m e believed that because memory was a (wax) reproduction o f an original event, its correct order was maintained. That is, a person w i t h a good memory for a cricket match recalls not only all the various events wi th in the match but also the order i n w h i c h they occurred. O n the other hand, the imagination is a process characterized by a higher level o f creativity. It can combine or break d o w n images and then re-arrange them, as i n the case o f poetry and literature (Copleston, 1964). Importantly, H u m e viewed the imagination as a vol i t ional and conscious process; created memories for false events (the modern conception) were not considered i n his writings. Treatise provided an insightful summation o f autobiographical memory qualities f rom the classical perspective. Despite flaws i n his view o n perception, some o f Hume ' s observations on imaginal processes have stood the test o f time and contribute to our current understanding o f the reconstructive nature o f imagination (under w h i c h created memories fall) and the qualities o f memories compared to imagination. Conceptions o f M e m o r y Dis tor t ion i n the 19th and Ear ly 20th Centuries O n e might expect that after the writings o f Le ibni tz , K a n t , and H u m e , the phi losophical mmunity w o u l d have devoted m u c h thought to problems o f memory. H o w e v e r , w i t h a few prions (e.g., A r t h u r Schopenhauer), few nineteenth century theorists devoted m u c h writ ing to either memory formation or distortion. Nonetheless, an early conception o f the false memory emerged i n the writings o f Samuel Dunnet t i n 1862. CO exce The Architecture of Childhood Memories 25 Dunnet t (1862) first describes false memories In L o n d o n i n 1862, an essay entided Phi losophy o f memory by Samuel Dunnet t appeared, litde k n o w n today, w h i c h may have been one fo the first usages o f the term "false m e m o r y " (p. 67), subsequendy adopted by W i l l i a m James. T o my knowledge, this paper provided the premier explicit exploration o f memory distortion. A c c o r d i n g to Dunnett , distortions had three primary origins: (1) lied-about events w h i c h became believed by the liar; (2) a weak w i l l w h i c h erroneously attributed truth to an imagined event, and; (3) additive distortion i n the context o f ordinary forgetting. In this essay, Dunnet t first detailed his thoughts o n h o w people decide whether a personal memory is based on a real or false event (often called "reality m o n i t o r i n g " today). H e felt that, i n general, we have litde prob lem distinguishing fact f rom imagination because a factual event has "been perceived, and felt and k n o w n " (p. 57) and becomes an established truth i n the soul. However , there are occasions w h e n a person is unsure whether an event happened, and may have self-interests i n assuming that i t d id . A c c o r d i n g to Dunnett , this situation rendered people susceptible to false memories, i n w h i c h they are " l e d to receive that for a real thing w h i c h i n fact had its origin i n nothing but the imaginat ion" (p. 58), and w h i c h can endure for decades. Dis tort ing influences on memories were often founded i n deception to others and to self, i n Dunnett ' s view. False memories seeded i n deception to others were called deceitful or treacherous memories. H e argued that i n most cases o f intentional fabrication o f an imagined event (i.e., lying), some faint traces o f truth may be found and that such cases are ultimately detectable: " A careful observer o f human conduct, by taking particular notice o f countenance, gestures, and frequentiy a degree o f embarrassment that attaches to the person" (p. 61) could successfully detect the treacherous memory (modern deception researcher Pau l E k m a n has spent the last thirty years attesting to this; e.g., E k m a n , The Architecture of Childhood Memories 26 1992, 1996). Nonetheless, w i t h the passage o f time aspects o f a treacherous memory originally seen as false by the person could come to be v iewed as true: Suppose, for example, that a circumstance occurred, say, several years ago, to w h i c h we were eye-witness, and suppose we thought at that time that it w o u l d serve our interest, to put some degree o f misconstruction o n it, and tried to tiiink there was no harm i n it, and suppose we commenced to ponder that misconstruction, and continued to revolve it i n the m i n d t i l l we thought that in that plausible form we might be comparatively safe i n telling it to others. Af ter telling it once i n this mutilated form, k n o w i n g it was not correct, yet, to appear consistent, when relating it again it must be to ld i n the same way. N o w , by this time it is reasonable to suppose, and a daily observation o f human conduct proves it, that the m i n d gains additional confidence i n its undertaking, and feels itself to be under stronger obligations to relate the circumstance the third time agreeable to the two former times; this must also be done to appear consistent.. .That is to say, the m i n d has for many years assumed improper modifications, w h i c h do not present the fact but a form w h i c h gready lessens its force. . . A n d , suppose that, under these circumstances, we should be called u p o n to give our testimony to that occurrence, m u c h as we might desire, at our present stage o f l i fe , . . .we can only give it as we have i t . . . in irregular form. In the absence o f mendacity, weak-willed individuals (recall the Greeks ' idea that memory problems o f the weak were due to deficient wax surfaces) could eventually believe w i t h confidence that an imagined event was an experienced event. Dunnett 's discussion o f the "honest forgetful m e m o r y " centered o n the idea that most details o f events people witness are eventually forgotten. Mis informat ion could be added to the memories to render them more plausible, but, for the most part, honest testimony w o u l d contain a poverty o f detail and litde The Architecture of Childhood Memories 27 coherence, i n contrast to the assertion o f H u m e . Dunnet t also painted a very different picture o f memory f r o m K a n t and L o c k e , the latter two believing that the power o f reason w o u l d generally al low us to accurately recall events. Dunnett 's memory images based o n " impress ions" could be inundated w i t h false sensory information founded i n deceptive ideas and language. Ear ly psychiatric and psychological views on memory distortion T h e origin o f m u c h thought o n memory construction came wi th in late nineteenth and twentieth century psychology and psychiatry, under the parentage o f existential philosophy. Af te r Dunnet t , one o f the first writers to devote extensive attention to memory distortion was eminent existential phi losopher/ psychiatrist K a r l Jaspers (1883-1969) i n his General psychopathology (trans. 1946). The subjective phenomena o f psychic life inc luding false perceptions and instances o f distorted memory attributable to abnormal imagery were elements o f Jaspers' system. Contending that all memories were subject to m i l d problems o f perceptual and temporal distortion, Jaspers also discussed extreme cases such as hallucinatory memories w h i c h are for false events but accompanied by a belief that what is remembered happened but had once been forgotten. This idea is remarkably similar to idea o f "recovered memories" , one o f the key issues i n the false memory debate (i.e., one " c a m p " i n the debate argues that most or all once forgotten but recovered chi ldhood memories are actually created memories; B r o w n et al., 1998). Ex i s t ing delusional ideas could confer new meaning to past events. F o r example, a persecutory delusion could lead one to remember past interpersonal interactions i n line w i t h the delusion, a c o m m o n thesis explored by the great existentialist, E d m u n d Husserl , soon after (see below). Jaspers can be credited as being an early observer o f everyday memory distortion. Severe distortions o f past events and deja v u - like experiences were also addressed i n his writings. The Architecture of Childhood Memories 28 Soon after W u n d t founded the discipline o f psychology i n 1879, the idea that memory is reconstructive and susceptive to distorting forces began to garner support. First, R i b o t (Diseases o f memory. 1882) emphasized memory's active and constructive nature and saw memory as occurring f rom a series o f reference points across time. Begmning i n Germany, the Aussage (German for "testimony") era o f applied research on memory distortion, initiated by French and G e r m a n psychologists (mainly Binet, Munsterberg, Stern, V o n Liszt , and Whipple ) , closely paralleled modern eyewitness memory research, methodologically and i n the nature o f the findings (see Y u i l l e , Daylen , Porter, & Marxsen, 1995). Suggestibility and the effects o f leading questions were important concerns during the early era (e.g., Binet, 1900; Stern, 1910, 1939). F o r example, Binet (1900) exposed children to various objects and then interviewed them w i t h or without misleading questions. Misleading information produced systematic distortions i n the memories, a f inding replicated wi th adults by Stern (1910). Seashore (1901), i n reviewing Binet's (1900) L a suggestibilite. shows the extent o f this early applied knowledge o f memory distortion: " H o w can a judge influence the witness by his questions? H e may allow the witness to exercise complete spontaneity; he may force questions without biasing; he may make leading suggestions to the witness by the questions" (p. 614). Indeed, opinions o n the constructive nature o f memory surfaced early i n forensic contexts (Table 2). T h e first recorded expert psychological evidence on memory distortion and reconstruction occurred i n 1896 (Bartol & Barto l , 1987). In this M u n i c h case, a suspected serial killer had murdered three w o m e n and Professor Alber t v o n Schrenck-Notz ing was called to give evidence as an expert. H e testified that the extensive pre-trial publicity had likely distorted the witnesses' memories through a process he called "retroactive memory falsification," to the extent that they could no longer be trusted. H e contended that the witnesses could not distinguish i n memory what they had seen happen and what they had heard i n the press (von Schrenck-Notzing's testimony was The Architecture of Childhood Memories 29 essentially ignored and the accused was convicted). Experts o n memory distortion also testified i n a 1911 Belgian case i n w h i c h someone had been serially murdering children. T h e only testimony implicat ing the accused came f rom chi ld witnesses. Varendonck , an expert psychological witness for the defense, provided the court w i t h a summary o f the Aussage research findings w h i c h had purportedly demonstrated the fallibility and malleability o f children's memories (see G o o d m a n , 1984), foreshadowing a sentiment echoed i n today's courtroom (e.g., C e c i & Bruck, 1993). In particular, he felt that children found it difficult to distinguish reality f rom imagination. O w i n g m u c h to Kant i an thought, Gestalt psychology i n the early twentieth century was a psychological approach i n w h i c h the interpretation and reconstruction o f perceptual experience was a premise. W u l f and other Gestaltists reported that i f research subjects were given " b a d " figures, such as incomplete geometric shapes, they w o u l d mentally fill i n the gaps to create " g o o d " complete figures. The perception o f ambiguous or i l lusional figures was largely interpretive and subjective. A s wel l , perception was viewed as holistic and not simply an additive process. O n o f the best-known theorists f rom the Gestalt school , K o f f k a (1935) reviewed an extensive body o f research o n h o w stimulus characteristics, context, and person variables create distortions i n perception such as the "ph i -phenomenon , " optical illusions, and the phantom l imb phenomenon and stated outright, "remembering appears to be far more decisively an affair o f construction rather than one o f mere reproduct ion" (p. 656). Recal l is not passive but is always trying to meet the needs o f the moment. Koffka ' s contemporary, famed neurologist Sir Henry H e a d (1920), stepped his prominent foot outside the door o f the storehouse armed w i t h his schema concept (Table 2). H e asserted that although the sensory cortex is the storehouse o f past impressions, the original impressions themselves were created not as the original sensory experience but by something he The Architecture of Childhood Memories 30 called a schema, essentially a modifying force resulting f rom past experience and knowledge. That is, schemata are ways o f l o o k i n g at the w o r l d to organize our past experiences and provide a framework for interpreting our future experiences, influencing bo th the perception o f and reconstruction o f an event. The idea o f schemata has been viewed as an important original contr ibution o f Head , as H e a d himself had mouthed. However , modern psychology may have given H e a d undue credit. A reference to schemata came 150 years earlier f rom Immanuel K a n t i n Crit ique o f Pure Reason, remarkably similar to the current conception i f one stumbles through his abstruse words. K a n t defined schema as "the pure synthesis determined by a rule o f unity, according to concepts, a synthesis as expressed by the category, and represents a transcendental product o f the imagination, a product w h i c h concerns the determination o f the internal sense i n general, under the conditions o f its fo rm (time), w i t h reference to all representations, so far as these are meant to be joined a prioriin one concept, according to the unity o f apperception" (p. 124). Perhaps his example is more intelligible: " T h e schema o f the triangle can exist nowhere but i n thought, and is i n fact a rule for the synthesis o f the imagination w i t h respect to pure forms i n space" (p. 123). A l t h o u g h reiterating Kant ' s thesis and terminology, H e a d appears to take credit for the idea by " p r o p o s i n g " the not ion o f schema (1920, p. 605). O r i g i n aside, the seminal schema concept was a very important concept i n developmental theories (e.g., Piaget) and facilitated a modern understanding o f memory. A r o u n d the same time that Head's pivotal thoughts emerged, A m e r i c a n philosopher J o h n Dewey's Reconstruct ion i n phi losophy (1920) was published. D e w e y advanced several notions about memory w h i c h correspond to a reconstructive view. First, he argued that revivals o f memory are rarely literal representations o f the original event but rather are more reflective o f the person remembering. Unintentionally, people recall what interests them most and what adds the most to the present circumstances. H e concluded that the primary nature o f memory is The Architecture of Childhood Memories 31 emotional or sentimental rather than intellectual and pragmatic, i n contrast to K a n t i a n thought. Further, D e w e y argued that the positive aspects o f events tend to be exaggerated and the negative aspects min imized , a process he called the idealising tendency, using the example o f a war batde: M e m o r y is vicarious experience i n w h i c h there is all the emotional values o f actual experience without its strains, vicissitudes and troubles. T h e t r iumph o f batde is even more poignant i n the memoria l war dance than at the moment o f victory; the conscious and truly human experience o f the chase comes when it is talked over and re-enacted by the campfire. A t the time, attention is taken up w i t h practical details and w i t h the strain o f uncertainty. O n l y later do the details compose into a story and fuse into a whole o f meaning...a drama emerges w i t h a beginning, a middle and a movement toward the cl imax o f achievement or defeat (p. 2-3). Dewey's idealizing tendency had been conceptualized earlier by many students o f memory not the least o f w h o m was W i l l i a m James i n considering the issue o f false memories: T h e most frequent source o f false memory is the accounts we give to others o f our experience. Such accounts we almost always make both more simple and more interesting than the truth (James, 1892, p. 217). W i t h i n psychoanalysis, the idealizing tendency was also an important concept. A s mentioned, a tenet o f this approach was that memory is spatial i n nature, that ch i ldhood memories can be stored as literal representations i n the unconscious. However , memory distortion served to facilitate the retention o f the painful memories i n the unconscious. A exemplar o f Dewey's idealizing tendency, Freud's "screen m e m o r y " was a recollection o f chi ldhood i n w h i c h the reality o f events was hidden behind a superficial, rose-colored screen The Architecture of Childhood Memories 32 image, distorted for self-protection (Freud, 1922). In modern psychology, the idealizing tendency has been witnessed i n many facets o f psychology. F o r example, H e n r y Bahrick and colleagues discovered that i f high school grades are not accurately recalled, distortions are usually inflations o f the grades (e.g., Bahrick & H a l l , 1996). There is also recent evidence that clinical depression may result f rom a failure o f the idealizing tendency, showing its importance i n our lives (e.g., M a c L e o d & Tata, 1997; M o f i t t & Singer, 1994) . Related to the idealizing tendency is the seminal social psychological theory o f "cognitive dissonance" (Festinger, 1957) by w h i c h people act or think to reduce the dissonance felt when two thoughts, i n a broad sense, are inconsistent (e.g., decision-making operates to reduce conflict between beliefs and perceptions). Dewey's third major contribution o n memory was to accentuate intra- and inter-personal factors i n distortion, noting that incidents wi th in an event are selected w h i c h have a present emotional value, to intensify the present tale as i t is rehearsed i n imagination or to ld to an admiring listener. In essense, he asserted that memory distortion serves to enhance a person's sense o f well-being or esteem, and that memory is largely a social psychological phenomenon. Bartlett refutes the storehouse view Buttressed by the m i n d o f Head , Sir Francis Bardett, a social psychologist, authored an important book o n memory i n 1932. W i t h Bardett's publication o f Remembering: A study i n experimental and social psychology, the reconstructivist perspective had its most significant empirical validation to that date. Rejecting outright the storehouse conception (and the spatial view generally), Bardett proposed the alternative v iew that remembering involves a reconstruction o f an event through the use o f pre-existing knowledge leading to the product ion o f a coherent narrative. H e postulated that the process o f remembering was inherendy a social The Architecture of Childhood Memories 33 activity distorted by the knowledge, attitudes, and personality o f the rememberer. Whereas K a n t had emphasized rationality and Dewey emotion i n h o w we remember and distort life events, Bartlett argued that memory is fundamentally social. Bardett exposed his subjects to an interesting o ld Indian legend called "The W a r o f the Ghos t s , " after w h i c h they were asked to retell the story o n multiple occasions. The retellings were never completely accurate. Participants often mentioned events w h i c h made intuitive sense or met their expectations o f what should have occurred. A s wel l , the memories o f the subjects evolved w i t h each retelling. Bardett used the not ion o f schemata to explain the pattern o f findings. People processed and recalled informat ion i n line w i t h their existing knowledge, beliefs, and social needs. Publ i shed during the reign o f behaviorism, Remembering d id not have a great impact o n psychology for decades but foreshadowed the ground-breaking work o n memory distortion i n the 1970s and 80s and is n o w regarded as a classic w o r k (see Schacter, 1995b). Behaviorists such as Watson and Skinner rejected the storehouse view dominating cognitive psychology (filing systems, computers, libraries, etc.) but, arguably, for the w r o n g reasons. Because o f its incompatibi l i ty w i t h the stimulus-response model , Skinner argued that, " the assumption o f a parallel inner record-keeping process adds nothing to our understanding o f this k i n d o f t h i n k i n g " (Skinner, 1974, p. 122). Inevitably, any modern reader o f behaviorism's explanatory version o f memory comes away wi th the sense that it was as vague as it was simplistic and, not unexpectedly, anti-mind. Behaviorists argued that memory was based not o n a search o f stored material but rather on. probability. Reinforcement influenced the frequency o f repetition o f a particular i tem or event increasing the probability o f remembering. T h e probability o f the "response" - the memory - was increased by being reminded o f the i tem initially and then re-experiencing it behaviorally (read perceptually). Curiously, there was litde r o o m for either a spatial or an architectural v iew f rom the behaviorist interpretation. N o storage, no imagery, no The Architecture of Childhood Memories 34 reason, n o social factors, no creativity - no mind . A l t h o u g h Bardett was easily dismissed i n this atmosphere, after Remembering, experiments i n social psychology soon began to support that memory is reconstructed. A wel l -known study by A l l p o r t and Postman (1945) indicated that memories can be altered by attitudinal factors. Participants were shown a picture o f a white male hold ing a razor as he argued w i t h a black male. U p o n questioning, 50% remembered the black male wie ld ing the razor and some recalled h i m "threatening" the white male. T h e importance o f personological variables, such as personality and defenses, i n perception and memory was recognized i n the " N e w L o o k " research o f the 1940s and 1950s by Jerome Bruner and colleagues, i n the Gestalt tradition. Exempl i fy ing this research was a study by Bruner and G o o d m a n (1947) i n w h i c h it was found that poverty-stricken children consistentiy overestimated the size o f coins compared to wealthy children but were proficient at judging the size o f cardboard disks. Similarly, hungry subjects were more likely to spot food i n ambiguous pictures. M a n y o f the familiar projective tests like the Rorschach and the Thematic Appercep t ion Test ( T A T ) had their origins i n the awareness that people process events according to their needs, schemata, o r "current mental structures" i n the philosophy o f E d m u n d Husser l (outlined below). A few theorists adopted the reconstruction perspective i n the early 1960's despite the Penfield-inspired resurrection o f the classical perspective. V o n L e y d o n (1961) described his " c h i l d h o o d test case" i n w h i c h Person A hears about an event o f Person A ' s ch i ldhood f rom Person B and Person A comes to reconstruct a memory for the event w h i c h is erroneously believed to be a real memory (as I mentioned near the beginning o f this paper, most people indeed report ch i ldhood memory conflicts wi th family members). T h e accuracy o f ch i ldhood memories were questioned by Smith (1966) w h o described the case i n w h i c h a person believes (s)he has an accurate memory o f a place f rom chi ldhood but upon returning finds a very The Architecture of Childhood Memories 35 different place, a phenomenon to w h i c h most o f us can certainly attest. The most radical proponent o f the reconstructive perspective may have been the originator o f the Weschler A d u l t Intelligence Scale (WAIS) D a v i d Weschler (1963), w h o offered that memories are not stored i n any manner: "memories do not exist...ideas are no more stored i n the brain than melodies i n the keys o f a p i a n o " (p. 150-151). Thus , the construction, or architecture, o f memory has been alluded to f rom early thinking o n memory, although usually i n deference to the dominant storehouse conception. Ar i s tode viewed " reco l lec t ion" as involv ing inference based o n a person's current situation and knowledge. Aristode's torch was handed by Leibni tz to K a n t w h o ran w i t h it to his little apartment i n Konigsberg . Rejecting strict storehouse notions, K a n t instead regarded memory as fundamentally dependent on active reasoning. Perception was subjective - mental frameworks or schemata interpreted the external wor ld . Intelligence, knowledge, and reasoning then served to re-create what a particular perceptual pattern was like, adopted by Gestalt psychology after the first W o r l d War . Dunnet t likely introduced the term false memory and discussed distorting factors o f deception. N o t only d id he discuss created memories, but also the phenomenological features o f ly ing w h i c h was not addressed again unti l Italian criminologist, Cesare L o m b r o s o , looked at physiological approaches to lie detection i n the late nineteenth century. T h e reconstructive nature o f memory was accorded experimental validation beginning i n the late nineteenth century w i t h Aussage research o n the effects o f leading informat ion on witness testimony. Forensic application o f reconstruction also took place early i n the form o f expert psychological testimony concerning the distortion o f eyewitness testimony. Dewey advanced several insightful ideas on event reconstruction, particularly his " ideal izing tendency," construed as screen memories by the psychoanalytic school. In 1932, Bardett snipped the remaining strand o f r ibbon and smashed the storehouse botde onto the ship Memory Reconstruction. H i s text left The Architecture of Childhood Memories 36 little question that memory had significant reconstructive aspects. Nonetheless, i n the reign o f behaviorism Remembering d id not exert m u c h influence o n memory theorists unt i l the late 1960s and 70s. A l t h o u g h this review reveals the acknowledgment o f reconstruction i n drinking o n memory, apart f rom D a v i d H u m e few writers had explored the phenomenological characteristics o f memory images. What is a reconstructed event like i n memory? W h a t is a highly distorted, or even imagined, event like i n memory? W e can speculate that K a n t w o u l d conclude that the memory image is primarily visual, coherent, temporally stable, and utterly "sensible" to the rememberer as a product o f reason. Dewey believed the details o f a memory image to be coherent and self-enhancing as a product o f the emotional self. Bardett w o u l d have thought that the memory image w o u l d contain scripted, expectancies - details i n accord w i t h the framework o f the individual's general outiook. Created memories were described by a few (e.g., Dunnet t , James) but the idea that a normal person could come to recall a completely false emotional event was not studied unt i l recendy. II. Reconstructive Views of Memory after 1967 Remember ing Remembering: Neisser's Reconstructive V i e w o f M e m o r y I n 1949, renowned Canadian psychologist D o n a l d H e b b discussed the inferential nature o f memory, drawing the parallel o f a paleontologist generating a complete dinosaur f rom bone fragments and a general knowledge o f dinosaur anatomy. Y e t constructive views o f memory did not become dominant unt i l the 1970s and 80s, fo l lowing U k i c h Neisser's influential Cognit ive Psychology (1967), and after the findings o f Penfield's studies were refuted. B o r r o w i n g Hebb' s analogy, Neisser's thesis was that memory for a life event was not simply a matter o f re-awakening a memory trace but rather involved reconstructing the event f rom pre-existing informat ion or general knowledge w i t h specific fragments o f knowledge o f what transpired The Architecture of Childhood Memories 37 (Table 2). H e concluded that the reconstructive nature o f memory left it subject to inaccuracy and distortion but he later underscored that memory can be highly accurate by its thematic nature (e.g., Neisser, 1982). Neisser's monograph finally highlighted Bardett's ideas o n memory reconstruction f rom Remembering for a wide academic audience. It was well-read and provided a major impetus for research on memory construction and distortion. A l o n g w i t h E l izabeth Loftus , Neisser has continued to be a primary advocate o f the reconstructive nature o f memory since this important publication. Cont inu ing Ev idence for M e m o r y Reconstruction i n C o g n i t i v e / Forensic Psychology In the 1950s, many scientists concluded prematurely that Penfield's results w i t h epileptics validated the permanence-of-memory (storehouse) conception. Proponents o f the resurrected v iew eventually capitulated as damning information surfaced about Penfield's results (e.g., see K n a p p & Vandecreek, 1997; Loftus , 1993; M a l c o l m , 1977). F o r example, as conceded by Penfield, only a small minority (40/530) o f his patients reported memories on stimulation o f brain sites and there was no basis to determine the accuracy o f the memories. Other sources suggest that Penfield and others misrepresented the results o f the temporal lobe stimulation studies and failed to report that the stimulation o f the same lobe sites elicited different memories at different times (Malcolm, 1977). In the early 1950s, K a r l Lashley had expressed an anti-localizationist view - that memories must be distributed throughout the brain. Initially hop ing to discover where memories were stored, he trained rats to solve a maze, then excised various pieces o f the cortices and re-tested them o n memory for the maze. Regardless o f w h i c h part o f the cortex he removed, the rats retained at least a partial memory for the maze forcing Lashley to conclude that memories do not reside i n different spots or that different aspects o f memories are processed i n different regions, i n contrast to Penfield's conclusion. B y The Architecture of Childhood Memories 38 the time Neisser's (1967) monograph appeared, empirical support for the reconstructive aspect o f memory had already been accumulating, such as f rom the N e w L o o k school. Reminiscent o f the 1940s social psychology experiments, Sulin and D o o l i n g (1974) had participants read passages about a disorderly and undisciplined girl ; some were informed that she was H e l e n K e l l e r and others that she was C a r o l Harris . A week later, the first group members were more likely to recall reading the sentence, "She was deaf, dumb, and b l i n d , " although this informat ion had not existed i n the initial passage. Similarly, participants w h o had been asked to read a passage about a dictator were more likely to recall the statement, " H e hated the Jews particularly and persecuted them, " i f they had been reading about a fictitious A d o l f H i d e r than a Gera ld Mar t in . Such studies continued to demonstrate the unyielding presence o f human inference i n the task o f remembering. T h e early eyewitness researchers, as wel l as Bardett and Neisser, had discussed the ordinary presence o f distortion i n human memory. In the theater o f memory, Distortion took the lead under the direction o f El izabeth Loftus . W i t h the genesis o f the n o w wel l -known "post-event mi s in format ion" paradigm and the numerous studies generated by it (e.g., Loftus , 1975; Loftus & Davies , 1984; Loftus & Palmer, 1974; Loftus & Z a n n i , 1975; Loftus , M i l l e r , & Burns, 1978; Loftus , Schooler, & Wagenaar, 1985), memory came to be v iewed as fundamentally fallible and malleable. Similar to the methodologies used by Aussage researchers and Bardett, the mis information paradigm showed that memory was "especially fragile i n the face o f subsequent events" (Loftus, Fe ldman, & Dashiel l , 1995, p . 47), a phenomenon k n o w n as retroactive interference. Us ing creative experiments, Loftus and colleagues clearly demonstrated that some people can readily incorporate false information provided by external sources (such as an experimenter) into their memories for observed events (e.g., Loftus et al., 1995). In the original paradigm, the participants were shown a slide or video presentation depicting a crime or The Architecture of Childhood Memories 39 accident. A f t e r seeing it, participants were asked a series o f questions about particular aspects o f the event. T h e experimental witnesses received some misleading informat ion wi th in the questionnaire after w h i c h they responded to a series o f inquiries about the event. T h e critical concern was the extent to w h i c h the experimental witnesses include the misleading informat ion into their accounts o f what transpired i n the presentation. F o r example, some studies have employed a slide sequence o f an automobile accident as the memory stimulus (e.g., Lof tus et al., 1978). Exper imenta l witnesses were provided misleading information such as, " D i d the car stop at the stop sign?," w h e n i n reality there was n o stop sign but rather a yield sign. O n the subsequent recognition test, witnesses were asked to choose between a stop sign and a yield sign. M i s l e d witnesses chose the correct slide only 4 1 % o f the time compared to 74% by the contro l witnesses. In other studies, people recalling speed after v iewing an automobile accident made higher estimates w h e n adjectives like " smashed" were incorporated into the question as opposed to " c o l l i d e d " or " h i t " (Loftus & Palmer, 1974) and i n response to questions containing a definite article (did you see " t h e " broken license plate?) produced more false recognitions than questions w i t h an indefinite article such as " a " (Loftus & Z a n n i , 1975). M o d e r n eyewitness researchers have also re-discovered f rom Stern and the Aussage researchers hypermnesia, i n w h i c h progressively more informat ion about an event gets recalled w i t h successive recall attempts (e.g., Scrivner & Safer, 1988; Turde & Yui l l e , 1994). This f inding has been interpreted f rom an associationist perspective, and the encoding-specificity principle outlined by T u l v i n g & T h o m p s o n (1973); details can act as facilitating cues for other details encoded i n the same context (e.g. Geise lman & Fisher, 1989). Alternatively, relating back to Bardett's (1932) original conclusion, hypermnesia could represent additive reconstructive processes. This research i n the modern eyewitness era provided evidence that, at least i n control led contexts, memory could be influenced by subsequendy processed information, correct or The Architecture of Childhood Memories 40 incorrect. This paradigm has become a favorite o f eyewitness researchers but the meaning and relevance o f the effects remain under dispute (e.g., M c C l o s k e y & Zaragosa, 1985). Consistendy, memory has been shown to be prone to misinformation i f the focus o f the informat ion is peripheral rather than central (e.g., Yui l l e , 1980), i f the aspect o f the event was not noticed at encoding (e.g., Schooler & Loftus, 1986), and i f the source o f the mis information is credible (e.g., Cec i , Ross, & Togl ia , 1987). Yui l l e at al. (1995) concluded, " I n short, this extensive research effort has demonstrated that witnesses may accept a reliable person's informat ion about something trivial i n an event i f the witnesses didn't notice it at the time" (p. 1272). Unreso lved is the issue o f whether a witness' memory is amended or appended (changed or added to) fo l lowing the incorporat ion o f the erroneous information, although the very recent created memory experiments indicate that people can come to recall entirely false events. Some theorists, especially Loftus (e.g., 1983), have argued that the findings f rom controlled, experimental contexts reveal a fundamental quality o f memory - fallibility - and thus extrapolate to "real l i fe , " whereas others believe that the dynamics o f forensic situations gready lessen susceptibility to mis information (Bregman & McAl l i s ter , 1982; D o d d & Bradshaw, 1980). Thi s latter idea was borne out i n the results f rom studies o f witnesses i n real w o r l d contexts, such as Yu i l l e and Cutshall's (1986) study w h i c h found that witnesses to an actual shooting were able to accurately reconstruct the event and were highly resistant to leading questions about even peripheral aspects o f the event (compared to the laboratory findings) over a lengthy period. Nonetheless, these findings on memory distortion have prompted some researchers to conclude that human memory is "clearly unreliable" (Melchert, 1996, p. 438). There has been an emphasis o n the erroneous aspects o f autobiographical events, when, i n fact, the data o n memory for personally important events indicates that such events tend to be accurately reconstructed (e.g., Christiansen, 1992). The pervasive negative v iew o f memory malleability and fallibility emerged The Architecture of Childhood Memories 41 out o f the same paradigm largely because psychologists have not sufficiendy considered the generalizability issue (e.g., M c C l o s k e y & Egeth, 1983; Yui l l e , 1993; Y u i l l e & Wel l s , 1991). But w i l l a forensic witness be equally susceptible to false information to a laboratory witness? T h e distorting influences o f misinformation may be contextually-dependent or consequence-dependent (dependent o n the perceived ramifications o f erroneous decision-making). A n approach devised by H y m a n et al. (1995) uti l ized distortion for actual chi ldhood memories w h i c h appears to be a major innovat ion to deal w i t h the issue o f generalizability. The post-event misinformation and subsequent robust findings initiated an explosion o f research on the distortion o f memory (see Schacter, 1995a, 1995b). Evidence for reconstruction has emerged i n studies on memory for various things, including w o r d lists (e.g., Roediger & M c D e r m o t t , 1996), songs (e.g., H y m a n & R u b i n , 1990), and, o f greater relevance here, autobiographical events (e.g., Barclay & D e C o o k e , 1988; Bruck, H e m b r o o k e , & C e c i , 1997; H y m a n et al., 1995; H y m a n & Pendand, 1996). T h e mis information paradigm contributed gready to understanding memory distortion i n controlled contexts and has also generated field research o n misinformation effects (e.g., Yui l l e & Cutshall , 1986). In the past few years, it has led to the development o f different approaches to examine extreme distortion o f memory for life events. Studies o f such phenomena (including the present study) are k n o w n as "false m e m o r y " studies (see Table 3). A s described later, although researchers have been successful at implanting memories for events w h i c h never occurred i n a very small number o f studies to this point (Pope, 1996), the categorical nature o f the memories w h i c h can be implanted is u n k n o w n (e.g., Pezdek & Roe , 1997; Porter & Marxsen, 1998). Further, the mis information paradigm has highlighted that memories can be "di s torted" i n the sense that information not present at perception can be incorporated into a reconstructed event. W h a t might the results o f the mis information experiments suggest about the qualities o f autobiographical memories? A t a The Architecture of Childhood Memories 42 speculative level, it suggests that erroneous verbal information can become transformed into visual informat ion i n the memory image, as Dunnet t and Bartlett had earlier claimed, so that the image reflects this information superimposed onto or assimilated into the original perceptual information. Table 2 summarizes the major historical figures who have contributed to science's understanding o f memory distortion and the idea that memory is constructive or architectural. I n the next section, a new reconstructive perspective is offered. Table 2 Architectural Conceptions o f M e m o r y : Ma jor His tor ica l Figures and their Contr ibut ions His tor ica l F i g u r e / Perspective Contr ibut ion Aris tode (c. 335 B C / 1908) H e distinguished memory and recollection, v iewing recollection as an inferential investigation. Plotinus (c. 244 A D / 1962) H e was the first to disagree w i t h Aristotel ian wax view, viewing memory as active, not passive. Le ibni tz (1704/1908) H e rejected spatial views o f the Greeks and Augustine. M e m o r y is the result o f active reasoning and interpretation. Social information could help " f i l l i n the gaps." Immanuel K a n t (1781/ 1966) H e elaborated Le ibni tz ' idea o f rational memory. A l l mental faculties (e.g., perception, memory) were active and based o n current knowledge and the "product ive imaginat ion. " Samuel Dunnet t (1862) K a r l Jaspers (trans. 1946) R ibot (1882) Aussage researchers (c. 1895-1912) Professor A . v o n Schrenck-Notz ing (1896); Varendonck (1911) Associat ionism Psychoanalysts (e.g., Freud, 1922) Gestaltists (e.g., K o f f k a , 1935) Sir H e n r y H e a d (1920) The Architecture of Childhood Memories 43 H e introduced the not ion o f " schemata" (despite Head's subsequent claim) and foreshadowed tenets o f Gestalt Psychology. H e first described "false memories." Distort ions were due to misconstructions. A l l memories are subject to perceptual and temporal distortion. H e discussed severely distorted memories as the result o f delusions and hallucinations. Memories are reconstructed f rom a series o f temporal points. E x a m i n e d distorting effects o f leading questions, suggestibility, o n eyewitness memory. First expert witnesses on memory distortion ("retroactive memory falsification", suggestibility o f children, respectively). M e m o r y is founded o n Aristotel ian principles. Af ter a stimulus has entered L T M storage, memory failures are actually retrieval failures. They discussed "screen memories" as protective distortions. They examined perceptual and memoria l distortions. M e m o r y was not passive or objective but reflects "exigencies o f the moment . " A l t h o u g h adhering to a spatial view o f memory, H e a d evoked the schema not ion to explain perception and The Architecture of Childhood Memories 44 Bardett (1932) memory. Publ i shed Remembering i n w h i c h spatial views were rejected outright. Argued that remembering is a social activity and that events are reconstructed through knowledge, attitudes, personality, and dynamics o f social interaction (applied schema theory to memory). E d m u n d Husser l (1929, 1966/ 1974) Th i s founding father o f phenomenology submitted that memory for events represents a fundamentally new k ind o f consciousness. H e introduced the concept o f "current mental structures" - the past can only be experienced i n light o f current possibilities o f meaning. Construct ion o f details for past events are additive, subtractive, or dynamic because o f a current state o f being (knowledge, emotion, reasoning). Thus , each time we recall an event, the image is fundamentally different, although often accurate. H e considered memories for ch i ldhood to be especially subject to distortion because o f temporal perspective change. U l r i c h Neisser (1967) W i t h the publication o f Cognit ive Psychology. Neisser underscored Bardett's v iew o f memory as reconstructive and refuted Penfield's ideas. L i k e D o n a l d H e b b , Neisser viewed the process o f remembering as similar to a paleontologist constructing a dinosaur f rom fossil fragments and subject to distortion. T h e M o d e r n eyewitness e ra / T h e emergence o f the post-event mis information paradigm. The Architecture of Childhood Memories 45 Mis informat ion paradigm It was demonstrated that memory reports o f laboratory witnesses could be markedly influenced by erroneous information presented after seeing an event. M a n y o f the issues originally examined i n Aussage era were re-examined and similar results rendered. The False M e m o r y research era Research addressing false or created memories for entirely (1990s) false events i n normal populations (e.g., Loftus & Pickre l l , 1995) A n Existential Reconstructivist Perspective G a z i n g at the skeleton o f the backyard tree, I remembered the searing words she had not spoken, Reached up a pale hand i n a self-consoling plea, Answered not by a sweet stream o f tears, N o cheek for brief rest or to cascade over! Peering through the l imbs o f my backyard tree, (Missing Overcoat drooped over once shoulders), I remembered the sounds she whispered i n her breath to me. Felt her caresses and mi lk and b lood , Retreat into real but see - I am again! T h e fo l lowing discussion offers a new framework for thinking about memory and memory distortion called existential reconstructivism w h i c h can be summarized i n simple terms as the view that humans tend to think about and remember events w i t h the greatest relevance to their well-being. Before going into more detail, the writings o f E d m u n d Husser l and h o w they relate to reconstructivism are discussed. The Architecture of Childhood Memories 46 Husserl's current mental structures T h e important pre-Socratic phi losopher Heraclitus (533-475 B C ) stated, " a l l things flow, nothing abides. Into the same river one cannot step twice" (Avey, 1954, p. 13). Heraclitus disagreed w i t h his contemporaries' attempts to reduce all things to a single permanent reality, flunking instead that a fundamental characteristic o f the universe is perpetual change. Because o f this quality, human sensory experience could not be trusted as literal. Consciousness and memories were ever-illusory, creating an in-the-moment existence i n man, a sentiment related to Kant ' s idea that nothing exists outside o f subjective experience. Heraclitus ' perspective had its rebirth i n the writings o f existential philosopher E d m u n d Husserl . E d m u n d Husserl , the founding father o f phenomenology (defined as " a self-critical methodology for reflexively examining and describing the l ived evidence w h i c h provides a crucial l ink i n our phi losophical and scientific understanding o f the w o r l d " [Reeder, 1986, p.l]) rejected any semblance o f the storehouse or passive wax notions and stated: "Reproduct ion is not, as H u m e and the sensualistic psychology since H u m e believe, something like a copy o f a perception or a weaker echo o f it, but a fundamentally new kind of consciousness" (Husserl, 1966/ 1974, p. 151). (Nonetheless, note that i n Treatise o n H u m a n Nature H u m e thoroughly discussed the idea that i n examining an experience i n a new and critical way changes the experience itself). Husser l (1929) observed that memory involves comprehending or inferring the past i n terms o f the structures o f meaning that exist at the current moment - a creative process. In accordance w i t h the principal o f intentionality, all drinking and consciousness is "perspectival" i n nature, that is any event, person, or relation exists not as an isolated entity (a "diing-in-itself ' i n K a n t i a n terminology), but f rom a special perspective activated by current mental structures, related to but more specific than Kant ' s schemata. The essence o f Husserlian philosophy is that the past can only be experienced in light of the current possibilities of meaning. Extending and elaborating the theory o f The Architecture of Childhood Memories 47 Leibni tz and K a n t , Husser l maintained that memory is dependent o n the creative capacities o f the m i n d and not merely a passive reproductive process. F o r example, a m a n thinks about a meeting he had wi th others and, i n light o f later events, he suddenly realizes that they were i n col lusion against l i i m (this is similar to Jaspers observation i n paranoid patients). H i s memory then changes to reflect this new knowledge/ belief and the man actually came to see the suspects signaling to one another and making devious moves, even though such details had not been noticed at the time o f event (Husserl, 1964). In a like manner, a detective wakes up i n the middle o f the night and realizes, " S m i t h committed the murder ! " H i s memory for the pertinent events during his investigation then becomes altered by this new knowledge. Thi s latter example also coincides w i t h Winson ' s (1997) recent conception o f (dreaming as the brain's efforts to re-consider past events, evaluate recent experiences, and plan new strategies to deal w i t h challenges i n the waking w o r l d . N o t e that i n Husserl's examples, the quality o f the memory changed considerably but the memory was not necessarily inaccurate i n theme. Current structures o f meaning enriched the memory w i t h additional information w h i c h made sense to the person at the time o f remembering. B u t the newly remembered information may or may not be an accurate reflection o f what occurred given that it became incorporated via a rational, logical decision-making process. Similarly, a subject i n a misinformation study might logically infer that the experimenter is provid ing accurate information and thus incorporate it, or a client might take a suggestion f rom a therapist and create conforming memory images. U n l i k e most earlier important thinkers o n this issue (perhaps w i t h the exception o f Freud and the psychoanalysts) such as H u m e , K a n t or Schopenhauer, Husser l considered the reconstruction o f memory for events o f the distant past (e.g., childhood) w h i c h he called "secondary memory , " using W i l l i a m James' term for long-term memory. M e m o r y for past events fades hastily " in to the v o i d " , The Architecture of Childhood Memories 48 according to Husser l (1964, p. 46) and must be recreated or "presentif ied." E a c h time a past event is recalled, a new perspective is brought to bear u p o n it and its quality w i l l be different (Husserl, 1966). H e asserted that memory and imagination were similar processes resulting i n a "phantasm" or pictorial mental image, except that a memory was based on the reconstruction o f an original perceptual impression and, further, that people have an awareness o f this (he does not appear to have explicidy considered the potential confusion o f completely imagined and real events). Experiences f rom the distant past were especially subject to memory distortion because o f inevitable changes i n knowledge and perspective, but could, as K a n t had maintained, remain accurate depictions o f what had occurred given opt imal conditions o f reasoning and interpretation. T h e quality o f the image itself w i l l be contingent o n w h e n the person is recalling an event. Husserl's explanation for memory and its potential distortion was r i ch compared to the modern post-event misinformation paradigm, w h i c h has lacked "theoretical structure or integration" (Yuille, 1980, p. 336). Clearly, Husser l the existential philosopher and Bardett the pioneering social psychologist held complementary views o f memory reconstruction. Ye t , as contemporaries apparendy they were either unaware o f or ignored each other's work. A l t h o u g h Husserl , a G e r m a n , began his important works early i n this century, he received scant attention f rom m u c h o f European phi losophy/ psychology unti l the 1950s. F o r example, Bertrand Russell's (1945) A history o f Western philosophy makes no references to Husserl . O n the other hand, G e r m a n philosophers recognized his insight earlier (e.g., M a r t i n Heidegger dedicated Be ing and time to Husser l i n 1926). In any case, Husser l first advanced his ideas o n memory decades before Remembering but the ideas, instead o f remaining o n phi losophical and speculative ground, were accorded early validation i n the empirical data o f the Aussage researchers and especially Bardett. The Architecture of Childhood Memories 49 A new perspective I do not recognize memory i n the sense that you mean it. Whatever we encounter that is great, beautiful, significant, need not be remembered f rom the outside; need not be hunted up and la id h o l d of, as i t were. Rather, f rom the beginning, it must be w o v e n into the fabric o f our inmost self, must become one w i t h it , create a new and better self i n us....There is n o past that one is al lowed to long for. There is only the eternally new, growing f rom the enlarged elements o f our past. - Goethe T h e present v iew holds that emotional memories are not stored as mental copies o f events but instead are architectural and constructive. T h e experience o f remembering an autobiographical event w i l l likely reflect the person's current circumstance and psychological state. That is, h o w the person perceives an event i n the first place w i l l depend o n h i s / h e r cognitive frameworks, or schemata, and personality. F o r example, the existence o f some people is abstract and prone to dissociation (see Bernstein & Putnam, 1986) whereas for others it is more concrete. These different ways o f look ing at and interacting w i t h the w o r l d may be i n some ways temporally consistent, but also continuously evolving. They w i l l likely influence perception, memory, and, possibly, susceptibility to false suggestions and created memories. Reconstructing the emotional event appears to be generally a rational process i n w h i c h current knowledge (i.e., general knowledge, knowledge o f the characteristics o f events, and knowledge o f h o w a particular event has affected a person) is drawn upon, w i t h an assessment (conscious or unconscious) o f h o w pertinent he reconstruction is to the person. F r o m this informat ion an image is re-created. G i v e n that this knowledge is ever-changing, each time an event is recalled the memory image and memory experience w i l l , to a certain degree, be altered. Thus , the objective reality o f the original events can be recreated w i t h varying degrees o f completeness and precision (e.g., K o r i a t & G o l d s m i t h , 1996; Wagenaar, 1996), but never perfecdy precise. The Architecture of Childhood Memories 50 A c c o r d i n g to the "remarkable memory" concept (Yuille & Cutshall , 1989; Yu i l l e & Daylen , 1997; Yu i l l e & Tollestrup, 1992; Yu i l l e et al., 1994), memories for certain personally significant events endure the test o f time, retaining a high degree o f clarity and detail. These events are thought about and remarked u p o n to others or oneself l imit ing the deterioration pattern characterizing some memories, as originally observed by Ebbinghaus, but increasing the opportunity for reconstructive distortion (especially peripheral information). Research o n memory for real and imagined events also suggests that tiiinking about and discussing personal events play an important role i n mamtaining the clarity o f the memories (e.g., Suengas & Johnson, 1988). T h e reconstructive nature o f autobiographical memories allows for greater integration and organization o f the numerous events we experience (Paris, 1996a). Schechtman (1994) pointed out that i f memory is to provide us w i t h a useful source o f knowledge about our o w n histories, we w i l l need to condense the information we receive. Because memories for events are not stored as originally encoded, the capacity o f the memory system is enhanced. That is, i f memory operated like a storehouse, it w o u l d presumably be able to " s tore" a certain number o f events, & finite capacity, w h i c h Penfield d id not consider. However , we often condense or expand an experience, providing ourselves w i t h a fictionalized event w h i c h may be a composite o f several real ones or w h i c h fits our knowledge base. A s wel l , humans may come prepared to " focus" o n certain events to the exclusion o f others. The events to be focused o n are those having the most relevance to the person i n a particular context. A likely candidate is the event w h i c h evokes a strong emotional reaction. C h i l d h o o d events o f an emotional nature generally have existential pertinence (e.g., such events may be important to safety) and may be reconstructed often over the lifespan. The Architecture of Childhood Memories 51 Despite the i l lus ion o f coherence, our moment-to-moment existence can be a complex chaos o f thought. A s James Joyce iUuminated i n his prose (e.g., the character o f young Stephen Daedelus i n Portrait o f the artist as a young man), mental experience is largely embroi led w i t h the senses, ideas, and emotions. F r o m this information, humans must draw a coherent picture o f the things most relevant to their lives. This may be accomplished through appraisal and reflection, fueled by the reconstructive forces o f reason and creativity. B o t h existing memories and knowledge (proactive influences) and new events (retroactive influences) appear to contribute to the reconstruction o f particular memories, or as Spencer and F l i n (1990) put it, "the ability to remember information is a function o f not only memory capacity, but also o f prior knowledge, mnemonic techniques, contextual cues, motivat ion, and emotional state" (p. 239). There are neurological activity correlates associated wi th recalling events but it is unlikely that certain memories are " s tored" i n certain brain regions (e.g., K o r i a t & G o l d s m i t h , 1996), although neural networks for types o f events may occur primarily i n brain regions. F o r example, as discussed later, some traumatic events appear to be sometimes recalled primarily wi th in areas o f the right hemisphere i n non-verbal fo rm (van der K o l k , 1997a, 1997b). Some research has indicated that imaginal processes involve the re-engagement o f the visual system. Payne et al. (1997) have hypothesized that the act o f remembering involves the "repercept ion" o f internal representations that are created f rom experiences w i t h the wor ld . This suggests reconstruction o f an event may be facilitated by sensory re-engagement rather than memories being " s tored. " The conclusions o f Bardett and Husser l have been upheld by the neuropsychological w o r k o f Gera ld E d e l m a n (1989, 1992) w h o has found that the brain is a constandy shifting active system i n w h i c h updating occurs i n light o f i n c o m i n g information. That is, the brain is a highly dynamic system i n w h i c h memories undergo continual modifications. The Architecture of Childhood Memories 52 Taken together, the data o n memories for life events indicate that memories are rarely factually precise and that include elements o f imaginative reconstruction i n w h i c h some degree o f distortion is normal (Paris, 1996a). This may create an erroneous portrayal o f autobiographical memory being completely distorted. However , autobiographical memory can be highly accurate without being a literal representation o f an event, as highlighted by Conway (1990): In autobiographical memory...it is not usually the case that a memory is completely false but rather that a memory relates to an event w h i c h did occur but not exacdy as remembered...it may be an important feature o f autobiographical memories that they are never true i n the sense that they are literal representations o f events, and i n this respect it makes litde sense to ask whether an autobiographical memory is true or false. Nevertheless, autobiographical memories may be accurate without being literal and may represent the personal meaning o f an event at the expense o f complete accuracy (p. 9). Thus , autobiographical memories are not true i n the sense o f being literal representations but instead are usually veridical i n theme w i t h some facts being preserved (e.g., Conway, 1990), particularly details central to the meaning o f the event. Neisser (1981) analyzed the now-famous case o f J o h n Dean's testimony i n the Watergate scandal. This case has been taken as an exemplar o f memory reconstruction. Neisser observed that an analysis o f Dean's reports indicated some instances o f memory evidencing the gist o f what was said o n a particular occasion. Elsewhere i n his testimony, however, there was litde correspondence between the course o f a conversation and his account o f it. A c c o r d i n g to Neisser, even i n those cases, however, there was usually a deeper level at w h i c h he was accurate: " H e gave an accurate portrayal o f the real situation, o f the actual characters and commitments o f the people he knew, and o f the events that lay behind the conversations he was trying to remember" (p. 4). In other The Architecture of Childhood Memories 53 words, memories for meaningful events can reveal closely the general nature o f what occurred but m u c h recollection occurs through inferential processes. Dean's memory, ostensibly impressive, was often erroneous i n the specifics o f meetings, dates, and statements. Thi s example is not as clear cut as has been assumed, however. There is no doubt that D e a n was "reconstruct ing" events, but the possibility o f deceptive reconstruction rather than sincerely erroneous reconstruction exists. That is, D e a n may have been recalling tilings i n a self-serving way remembering and "forgetting" information selectively (Yuille, personal communicat ion, 1996). I f he were a sophisticated liar, D e a n w o u l d have k n o w n that the provis ion o f many specific details could only enhance his credibility (e.g., B e l l & Loftus , 1988), an idea also recognized by Ol iver N o r t h during the Iran-contra scandal. A s wel l , slight distortions may enhance bekevability, recognized i n the maxim, " A litde inaccuracy sometimes saves tons o f explanations." Since reconstruction o f events entails creative and rational cognitive functions, there are implications for determining the validity o f a particular memory. It is possible that a person recalling a real event employs different patterns o f construction than w h e n fabricating an event and wishing to appear credible or when the person recalls a created memory. T h e question o f h o w humans construct their pasts is a fundamental one i n psychology. A l t h o u g h it can be durable and accurate, memory appears to be plastic and susceptive to both internal and external influences. This versatility, w h i c h o n first consideration appears to be a fallibility, has prima facie adaptive value over the course o f evolution, and may not be surprising given the plasticity o f the five senses (e.g., Teylor & Fountain, 1987) o n w h i c h all memories are based. T h e current perspective holds that i n order for any event to be transformed into an enduring, v i v i d memory it must be thought about a n d / or spoken about often and at a deep level and that this w i l l be especially characteristic o f emotional events. The events resulting i n enduring memories w i l l be those for w h i c h it is i n our best interest to communicate about them The Architecture of Childhood Memories 54 to the self a n d / or others. Paradoxically, though, w i t h each successive reflection, subde distortions, f rom social and experiential sources, can occur resulting i n an enduring but not necessarily accurate memory. A c c o r d i n g to the existential reconstructivist view, there are two main reasons w h y events o f an emotional nature should have specific reconstructive features resulting i n enduring memories. First, emotional events are generally more pertinent to people's well-being than non-emotional events and reflection is adaptive. Secondly, this m o d e l operates on the principle that through natural selection humans developed a need for reassurance that they exist and act u p o n their environments to receive maximal existential reassurance. Reflection o n particular events can satisfy this need better than others. E m o t i o n a l events also tend to threaten existence more than others and require more consideration. Some assumptions and broad foundations o f this existential reconstructivist v iew are presented below: 1. E m o t i o n serves a communicative function E m o t i o n a l memories, or memories for events accompanied by a strong emotional response, are more likely to be reflected u p o n often i n a person's life and to be recalled vividly. E m o t i o n serves to communicate to oneself and others the event's existential relevance (relevance to one's well-being). G i v e n the cognitive and cultural advances o f modern humans, existence encompasses well-being or quality o f l iv ing rather than only physical survival. A s Pi l lemer (1993) wrote, "the act o f recounting a detailed personal memory to others communicates meaning that transcends the surface content o f the particular recollection, and this specialized form o f communicat ion appears to be rule governed" (p. 236). The Architecture of Childhood Memories 55 2. M e m o r y functions according to an adaptation principle Evolut ionary principles have occasionally been applied to systems o f memory, as by R o z i n and colleagues (e.g., R o z i n , 1976; Sherry & Schacter, 1987), w h o observed that memory and learning abilities are adaptive specializations, shaped by natural selection to solve problems wi th in particular environmental conditions (Rozin, 1976; R o z i n & Kalat , 1971). It can be predicted w i t h some confidence that experiencing a nove l emotional event w i l l elicit primitive adaptive strategies, w i t h implications for memory, w h i c h have been selected for evolutionary advantage. Af ter the event, a person's private and shared reflection o n the event, although subject to a host o f factors, is likely contingent o n h o w pertinent the reflection is to h i s / her well-being. A s noted by van der K o l k (1996), it appears that evolution favors the recall o f personally important informat ion or events w h i c h have the most relevance to our existence. C o m m u n i c a t i o n to self and others w i l l occur often for events w h e n it enhances well-being and reproductive fitness o f self and others. In general, litde communicat ion w i l l occur for events i f there is no adaptive benefit for doing so. The most frequendy recalled memories w i l l be for events i n w h i c h the person's existence or another person's existence had been threatened. A s a consequence, negatively-valenced chi ldhood events, such as injury or v ict imizat ion memories, may be recalled more frequendy than positively-valenced events (Bruhn, 1995). M o s t v i v i d memories w i l l involve themes o f existential or reproductive potential. F o r example, people most vividly remember an injury, an accident, Ufe-threatening event, hearing about a death, a medical procedure, first sexual experience, a wedding, a risky but thril l ing experience, a divorce, the bir th o f a sibling, and so on . 3. Created/ implanted memories can also be adaptive It may be adaptive for a person to believe events discussed by family members or other trusted people, even i f the person cannot initially recall the event. I f someone is in formed by a The Architecture of Childhood Memories 56 trusted, authoritative person that an emotional event occurred to them it is i n that person's best interest to try to come to recall the event, event though the trusted source could be mistaken. In particular, it might be i n the person's best interest to try to recall an emotional event, i n line w i t h the communicat ive function. It is theoretically possible for a person to come to generate a memory image for the event, regardless o f its validity. The l ikel ihood that someone w i l l actually produce the desired memory may be based o n personality or cognitive features, such as capacity to dissociate or creative imagination (e.g., H y m a n & Bill ings, 1998; H y m a n & Pendand, 1996). 4. Memories for emotional events are generally enduring Despite a c o m m o n view o f memory as unreliable, memories for events w h i c h elicited an emotional reaction are generally expected to be recalled accurately. M e m o r y for emotional events is simultaneously fragile and powerful , often "ephemeral and distorted" but also mosdy accurate and enduring (Schacter, 1995b, p. 20). There is m u c h evidence that highly stressful and personally relevant memories are very well-retained (e.g., Christ ianson & Loftus , 1991; Wagenaar & Groenewed, 1990; Yu i l l e & CutshaU, 1986). 5. M e m o r y for life events is architectural and inferential T h e construction o f a memory for any life event involves an inquiry guided by inference o f h o w the event has affected the person and h o w it is best dealt w i t h for the person's wel l-being (from forgetting to v i v i d image maintenance). A n example is Dewey's " ideal iz ing tendency" or the tendency to recall and report aspects o f events w h i c h enhance rather than devalue the rememberer. It is not particularly meaningful to discuss "storage" o f the original event image. T h e memory image can be transformed f rom semantic/ verbal fo rm to a sensory image form and may contain elements o f actual experiences. The construction o f a memory image involves not only perception but human knowledge, reasoning, and interpretation. Reconstruction involves an active inquiry for what occurred based on current status or h o w we The Architecture of Childhood Memories 57 are at a particular time. A s Heidegger (1964) observed: " E v e r y inquiry is a seeking. E v e r y seeking gets guided beforehand by what is sought. Inquiry is a cognizant seeking for an entity bo th w i t h regard to the fact that it is and w i t h regard to its being as it i s " (p. 24). 6. Real, created, and fabricated memories involve the construction o f a memory image Created memories are similar to " r e a l " memories i n that for both , reconstruction is based o n l imited source information and current knowledge structures but the nature o f the source informat ion is different. F o r real ch i ldhood emotional events being recalled i n adulthood, the person likely recalled the event o n many occasions across the lifespan whereas for created memories, the person may have begun recalling the event only i n adulthood. In the case o f real memories, the l imited informat ion includes the initial sensory experience and a host o f internal and external factors. T h e l imited information on w h i c h created memories are founded sometimes includes an external suggestion (e.g., Loftus , 1997a). A n internal motivat ion to recall a suggested event can facilitate the creation o f an enduring, v i v i d memory for a false event. Consequential fabricated (lied-about) memories, such as those presented i n forensic contexts, can contain a memory image and are constructed for adaptive purposes (to protect the person's well-being). The subjective characteristics and content o f the memory report may be different f rom one based o n a real event. In summary, after 1967, eyewitness and social psychological research provided consistent evidence that mis information can be incorporated into memories for laboratory events. Together w i t h the other evidence discussed, it is clear that there are reconstructive aspects to autobiographical memories and that the original perceptual experience is probably not reinstated. M e m o r y can be malleable as evidenced by the effects o f mis information (e.g., suggestive questions), but does not appear to be as fallible as is commonly impl ied . T h e writings The Architecture of Childhood Memories 58 o f E d m u n d Husser l were applicable to the interpretation o f such findings. Basically, Husser l asserted that because perception "fades quick ly" into nothingness, the long-lasting effects o f perceptual experiences on a person's t l i inking must be used to remember events. Current structures o f meaning drive people's interpretation o f past events or what "must have occurred" given what we currendy understand to be true. M o d e r n memory theorists recognize that memory is not a literal representation o f an original event but that usually the gist o f important personal events is recalled accurately. Inaccuracies can occur due to interpersonal and private reflection o n an event. Mult ip le reconstructions, w h i c h w i l l occur for events appraised as having the greatest personal relevance, can enhance the endurance (retention o f accuracy over time) o f central details o f an autobiographical memory (a person w i l l k n o w what happened) while distortions associated w i t h evolving current knowledge w i l l be introduced w h i c h each reflection (but not exacdy what happened). Certain ch i ldhood events w i l l l ikely be recalled more often and more vividly than most and possibly negative emotional events more so than positive. F o r example, L o r d Byron's earliest clear recollections were o f his parents' violent fights, the corporal disciplines o f his nurse, and the fear he associated w i t h them (Russell, 1945), a phenomenon to w h i c h many abused as children can attest. Nonetheless, the same rational process w h i c h serves to remember important events wel l can make significant mistakes. T h e emotional impact o f certain events appears to motivate people to recall or reconstruct such events more carefully than others. O n some occasions, people rely o n the information o f other trusted people, notably their families but other authority figures including teachers, therapists, and so on. Sometimes, i n certain contexts, it can be i n a person's best interest to recall an event, even i f that person d id not originally experience it as significant a n d / or is initially unable to recall it. F o r example, a father, not realizing that it was actually his other son, repeatedly tells the w r o n g adult son that w h e n that son was five years o f age he was The Architecture of Childhood Memories 59 attacked by a G e r m a n Sheppard dog. It might be i n the son's interest to try to come to eventually recall the event, so that he can (for example) take extra caution w i t h such dogs, come to understand his fear around dogs generally, reduce the dissonance resulting f rom his memory failure, and so on . M e m o r y imagery reinforced by the father can serve to resolve the memory failure. T h e son being falsely in formed might have personality or cognitive features w h i c h either facilitate or reduce the l ikel ihood that a created memory is formed (cf. H y m a n & Bil l ings, 1998; H y m a n & Pendand, 1996). In summary, sometimes the information about an important event given to a person by a trusted source may be incorrect but it could be advantageous for the person to recall it anyway. Therefore, o n occasion, people can reconstruct wrongly memories for central aspects o f emotional events, as i n the personal anecdote I gave early i n the paper, or even for entire false events. In the case o f the K i l l e r Tomato attack, I might hypothesize that over the course o f many retellings i n the presence and absence o f m y friend, a series o f factors contributed to my erroneous identification o f myself as the v ic t im. N o doubt she received heartfelt pity f rom the audience i n the midst o f describing a fascinating, funny, and scary experience. Such attention may have been attractive to my m i n d and m y o w n desire to relate a good personal tale (a la Dewey). A l s o , note that this event was the subject o f my dreams over the years so the possibility exists that my memory was distorted i n this context. Further, given that the friend was, at the time o f the event, my romantic partner, m y current mental structures may have been such that I had absolved myself o f responsibility for the relationship breakup through rationalization, v iewing myself as v ict imized i n the relationship and i n our shared negative experiences. O v e r time, I was transformed f rom the bystander to the v i c t i m i n m y memory image, although all other central details o f the event remained consistent w i t h her version. M y confidence was very high i n my vict imizat ion and the whole event is crystal clear i n The Architecture of Childhood Memories 60 my memory. This image w i l l likely change i n the future as my perspective changes; even writ ing about the event could be altering aspects o f m y memory for it. Inaccuracies can also be intentional as i n lying, another form o f event reconstruction, w h i c h w i l l be discussed i n depth later. A n existential reconstructivist approach was outlined to guide thinking o n h o w to differentiate reconstructions based o n real and false experiences. W e can often confidendy guess the periods i n an artist's life when different paintings were created even though they had the same creator. Similarly, we might detect the validity o f event reconstructions based o n a person's presentation style, or the content o f the memory report. Next , some theories applied to memory for emotional events w h i c h still exert m u c h influence wi th in psychology (such as " f lashbulb" memories) are reviewed, and possible characteristics o f real, created, and fabricated memories are addressed. T o conclude this discourse o n this modern view o f the reconstructive nature o f memory, a quote f rom Ari s tode (c. 350 B C / 1908) shows his awareness o f an architect residing i n the mind's storehouse : T h e act o f recollecting differs f rom that o f remembering, not only chronologically, but also i n this, that many also o f the other animals [as wel l as man] have memory, but, o f all that we are acquainted wi th , none we venture to say, except man, shares i n the faculty o f recollection. T h e cause o f this is that recollection is, at is were, a mode o f inference. F o r he w h o endeavors to recollect infers that he formerly saw, or heard, or had some such experience, and the process by w h i c h he succeeds i n recollecting is, as it were, a sort o f investigation (p. 71). The Architecture of Childhood Memories CHAPTER THREE Notable Returns to the Storehouse The Architecture of Childhood Memories 62 Travels Back to the Storehouse A l t h o u g h reconstruction is gaining acceptance as a mode l o f h o w memory for life events operates, psychology continues to make notable sojourns back to the storehouse. Storehouse views still exert influence i n the discipline, as we l l as the courtroom, and it is necessary to weigh the evidence for some o f the more dominant storehouse conceptions. A n overview is provided o n some o f the more popular storehouse cases, inc luding flashbulb memories, repressed memories eidetic memories, and Luria's case o f S. "F lashbulb" memories T h e storehouse metaphor o f memory was resurrected by B r o w n and K u l i k (1977) i n their theory o f the flashbulb memory - a v i v i d and accurate recollection o f the context of, or the circumstances surrounding, a highly emotional cultural event such as hearing about a president's assassination. N o t i n g the subjective vividness o f and confidence i n the memories and apparendy assuming their accuracy, the authors concluded that these memories were mediated by a special " N o w Pr int ! " mechanism, resulting f rom natural selection, w h i c h "freezes" the exact details o f a scene i n one's m i n d , similar i n effect to photographic, or eidetic, memory. Researchers inspired by this original paper amassed considerable findings i n support o f the flashbulb memory (e.g., Christianson, 1989; Winograd & Ki l l inger , 1983; C o h e n , M c C l o s k e y , & W i b l e , 1988; M c C l o s k e y , W i b l e , & C o h e n , 1988). Other studies have indicated that, although these events are v iv idly and confidendy recalled, they are often inaccurately recalled (e.g., B o h a n n o n & Symons, 1992; Neisser, & Harsch , 1992). Neisser and Harsch (1992) collected data on people's recall o f hearing o f the Challenger explosion wi th in 24 hours o f the event and 2.5 and 3 years later. They reported that only 7% o f participants showed perfect recall for the event wi th 2 5 % o f the reports being The Architecture of Childhood Memories 63 completely inaccurate. They took these results as evidence for the reconstructive v iew o f memory - that flashbulb memories, like other forms o f autobiographical memories, are distorted by schemata (e.g., Paris, 1996a). A d o p t i n g a similar approach, Christ ianson (1989) examined the consistency o f 36 Swedish participants' memories for hearing o f the assassination o f Swedish prime minister, O l a f Palme, over a one-year period. H e found that the event was accurately recalled i n terms o f a general narrative conception but specific details o f the circumstances were often erroneous prompt ing the conclusion that "these memories appear to be reconstructions based o n residuals o f the circumstances concomitant wi th the specific event (i.e., that o f hearing o f the shocking news), and these memories fo l low the same pattern o f recollection o f other autobiographical . . . emotional events" (p. 435). A personal diary study by Larsen (1992) supports Christianson's findings. Larsen recorded daily events for six months and tested his memory f rom one to eleven months later. Unexpectedly, during this period, two major events occurred - the assassination o f Palme and the Chernobyl disaster. Larsen's memory for both events was highly distorted. F o r example, he vividly recalled hearing the news about Chernoby l "at Pia's place just after w o r k " but according to his diary he had heard about it o n the radio while alone i n the m o r n i n g before going to work. In a study o f memory for the context o f Princess Diana's death (e.g., where the person was and w h o they were with) over a per iod o f six months, preliminary results suggest that the memories o f a considerable proport ion o f participants change over time, but confidence i n their memories does not (Porter & H e r v e , i n preparation). Accord ing ly , it appears that the "flashbulb memory" does not represent a unique class o f impeccable recollections but are rather a subset o f memories for highly emotional events generally (e.g., Brewer, 1992; Christianson, 1989, 1992). The " N o w P r i n t " explanation is reminiscent the Aristotel ian wax concept despite its intuitive appeal. Interpreted f rom the reconstructivist view, events such as the Challenger explosion are characterized by shock (for The Architecture of Childhood Memories 64 some people), frequent early reconstruction o f the init ial event, subsequent numerous exposures to the event, and numerous retellings to others. B o h a n n o n and Symons (1992) found that participants w h o reported the context o f the Challenger event inconsistendy over time also reported retelling their stories three times as often as other participants. A misleading term was selected by B r o w n and K u l i k (1977) i n "f lashbulb memories . " There do not appear to be mental "photographs" o f the original event but reconstructions w i t h varying degrees o f accuracy. There is no doubt that people can h o l d v i v i d memories for the events w h i c h researchers have defined as emotionally-arousing cultural events such the Challenger explosion, although researchers have not often assessed individuals ' appraisals o f such events. People may believe they possess mental photographs o f important events. T h e images can be sometimes subjectively v i v i d and enduring but, as we have seen, can be greatly distorted, as i n the case o f the K i l l e r T o m a t o attack. Pi l lemer (1984) found that the level o f affect present at the time the person heard the news was related to the sensory nature o f the memories, the consistency o f the memory over time, and h o w elaborate the memory was. The existential reconstructivist interpretation holds that people's potent affective experience at the time o f hearing o f such an event intimates its relevance to their existence. Depend ing o n the appraisal, a process o f private and social reflection o n the event can be spurred (whereas an ordinary life event w o u l d be appraised as having litde existential pertinence and the process w o u l d not be provoked). Relating to the reflection process is a reconstructed image characterized by vividness but also a degree o f distortion f rom both internal and external sources. H o w do memories containing the flashbulb quality differ f rom other memories? Brewer (1992) noted that flashbulbs often contain highly idiosyncratic details (e.g., " w e all had o n our litde uniforms") but this quality had not yet been compared to non-flashbulb memories. B r o w n The Architecture of Childhood Memories 65 and K u l i k (1977) had observed that flashbulb memories contained details o f place, affect o f self, and affect o f others. However , Brewer (1992) noted that even memory for ordinary life events contains these details (except possibly for reduced ment ion o f affect). A defining feature o f flashbulbs is their v i v i d quality; the images usually have clarity and are detailed. However , there can be content errors such as source distortions (where and f r o m w h o the person heard about the event), temporal distortions, and perspective distortions. Information central to the event tends to be reconstructed accurately so that, i n the case o f Challenger, the explosion image may be recalled vividly and accurately but information appraised as peripheral to a person's existence, such as the shutde's take-off, the personal context o f the event, and so o n may be recalled vividly but erroneously. O f course, a person recalling the context o f the Challenger news is not recalling an event w i t h the direct personal involvement characteristic o f many emotional memories w h i c h may have qualitative implications. Nonetheless, memories for emotional events f rom ch i ldhood containing this subjectively v i v i d , flashbulb quality may be operating under a similar principle as " f lashbulb" memories, maintained by private and interpersonal reflection. A g a i n , we recall what is most important f rom our o w n perspectives and i n accordance w i t h our current mental structures. O n a personal note, I h o l d a clear memory for hearing o f the Challenger disaster. However , i n the lucid image I can see myself lying o n my bed watching the news while talking on the telephone to my friend. This is an observer perspective w h i c h w i l l be addressed further as a possible credibility assessment tool . A s wel l , I a m aware that the v i v i d image o f the Challenger context changes. That is i n one earlier image, I could view the side profile o f my facial expression u p o n seeing the televised disaster but i n the current one, I can only see the back o f my shaking head. In terms o f the created/ recovered memory debate, it should be kept i n m i n d that although stressful or traumatic chi ldhood events may be recalled like "f lashbulbs" , they too are The Architecture of Childhood Memories 66 l ikely to have undergone some distortion and a flashbulb quality does not necessarily imply credibility. Some reported "recovered" memories, such as that o f E i l een Frankl in w h o claims to have recovered a ch i ldhood memory o f witnessing her father k i l l her friend, are reported to be recalled as virtual photographs coming back into the m i n d (MacLean, 1993). In Franklin's case, the flashbulb quality o f this memory was sufficiendy compell ing for jurors to convict her father. Some probable created memories have been characterized as being highly v i v i d (Lanning, 1991). T h e Repressed M e m o r y : A prevalent modern storehouse not ion Remembrance has a rear and front,-'Tis something l ike a house; It has a garret also F o r refuse and the mouse, - E m i l y D i c k i n s o n (1959) C a n a person hide unpleasant ch i ldhood events i n the recesses o f the unconscious? T h e concept o f repression is an important, perhaps the foremost, modern spatial no t ion o f memory. Part o f the r ich legacy o f psychoanalysis (see Ellenberger, 1970; Sulloway, 1992), repression is commonly evoked, often i n forensic settings, as an explanation for failures to recollect emotional events f rom chi ldhood. Variations o f the term repression have been around since at least the 1300s (much earlier, Socrates [369 B C / 1972 A D ] had considered the idea o f ho ld ing a memory but losing conscious awareness o f it). Represser; is thought to have its or igin i n the L a t i n and O l d French languages (repressus; represser) and had a number o f denotations i n M i d d l e E n g l i s h language including: (a) to overcome, put down, or suppress a riot, rebell ion, or disagreement; (b) to suppress sin or error; (c) to counteract v e n o m or pain; (d) to weaken i n effect or function and; (e) to reduce i n size (Middle Engl i sh Dict ionary, 1984). T h e second usage o f the term is most l ike the modern psychological version and, although it never referred direcdy to the suppression The Architecture of Childhood Memories 67 o f a memory (to m y knowledge), it d id sometimes concern the suppression o f immorality, especially sexual urges (the M i d d l e Engl i sh dictionary [1984] provides many cases including: " G o d . . . w o l represse and putte doune frome ye alle ye bestial desires" [c. 1456]). However , it appears that the historical observers o f human nature d id not witness cases o f repression, though Matthew A r n o l d once wrote " W e forget because we must, A n d not because we w i l l " (it is unclear whether he was referring to repression). A l t h o u g h the origin o f the c o m m o n idea o f repression concept has commonly been attributed to Freud, eighteenth and nineteenth century G e r m a n philosophers were among the first to consider the phenomenon. In his pr incipal work , The w o r l d as w i l l and idea (of w h i c h the author claimed that certain paragraphs had been dictated by the H o l y G h o s t [Russell, 1945]), Schopenhauer (1818/1896) outlines the weaknesses o f the human intellect, one o f w h i c h is the requirement that we store certain important thoughts i n the unconscious i n order to attend to more immediate ideas: Therefore after a certain time, w h i c h varies w i t h the individual , we must for the present give up every meditation or deliberation w h i c h has had the fortune to remain undisturbed, but yet has not been brought to an end, even i f i t concerns a matter w h i c h is most important and pertinent to us; and we must dismiss f rom our consciousness the subject w h i c h interests us so m u c h , however heavily our anxiety about it may weigh u p o n us, in order to occupy ourselves n o w wi th insignificant and indifferent things. D u r i n g this time that important subject no longer exists for us; it is like the heat i n co ld water, latent (p. 331). Schopenhauer asserted that such "dismissed" thoughts were retrievable but w o u l d influence future ideas because o f the physical composi t ion o f the humours and tensions o f the nerves change over time (in the Aristotel ian tradition, Schopenhauer believed they left an echo The Architecture of Childhood Memories 68 behind them, "the tone o f w h i c h influences the ideas w h i c h f o l l o w , " p. 331). H e noted that humans are able to unconsciously suppress, negative thoughts, observing, " i t is a curious fact that i n bad days we can very vividly recall the good time that is n o w no more; but that i n good days we have only a very co ld and imperfect memory o f the b a d " (1851/1951, p. 49), foreshadowing Dewey's idealizing tendency. N o t generally k n o w n as a man awash w i t h opt imism, Schopenhauer concluded that human memory is a fallible process, incomplete, and the litde that is recalled is inaccurate. H a l f a century later, Freud introduced his version o f repression, surpassed by few other concepts i n its influence o n psychological thought i n this century (Table 1). T h e influence o f repressed memories was a key foundation o f psychoanalysis, reflected i n the familiar declaration o f Breuer and Freud (1895/ 1955): "hysterics suffer mainly f rom reminiscences" (p. 7). A c c o r d i n g to F r e u d (1922), repression involves an emotional event being actively buried i n the unconscious, a process independent o f ordinary forgetting. Freud defined trauma i n terms o f the economy o f psychological energy expended during an event resulting f rom excessive stress or "stimulation.' 1 T h e traumatic event was one w h i c h "wi th in a very short space o f time subjects the m i n d to such a very high increase o f stimulation that assimilation or elaboration o f it can no longer be effected by normal means, so that lasting disturbances must result i n the distribution o f the available energy i n the m i n d " (p. 232). The memory remained intact but inaccessible because a "connection.. .had been broken" (p. 239). T o clarify the process, Freud used an explicit storehouse metaphor - an ante-room containing the unconscious and a smaller reception r o o m containing the conscious. O n the threshold o f the two rooms is a door-keeper w h o examines the various "mental excitations," censors them, and decides whether to admit them to the reception r o o m or the consciousness. Those excitations w h i c h are turned back by the doorkeeper are repressed (Freud, 1922). Thus , repression is a type o f amnesia i n w h i c h the The Architecture of Childhood Memories 69 psyche defends itself against intolerably painful knowledge. M e m o r y for a traumatic psychological event is not al lowed - via unconscious controls - into a person's conscious awareness. T h e modern psychodynamic view, like Freud's, is that such repressed material can continue to exert psychological effects o n the person over the lifespan and that a fundamental goal o f therapy is to recover the memory (Breuer & Freud, 1893/ 1957; Freud, 1922, 1974; Earleywine & G a n n , 1995; Myerson , 1977). Qualitatively, the recovered adulthood memory should be similar to the original memory at the time o f the original repression although there is no empirical support for this not ion (e.g., Loftus , 1993; Paris, 1996a). M a n y reports o f recovered memories o f abusive situations are characterized by a high degree o f detail and a "photograph-l ike" quality, although few cases have been substantiated. Near ly all cases o f reported recovered memories for Satanic ritual abuse, w h i c h generally contain such v i v i d qualities, remain unsubstantiated and most are highly improbable (Lanning, 1991). A m e r i c a n courts have occasionally convicted people o n the basis o f recovered, previously repressed, memories, exemplified i n the 1990 case o f George Frankl in . Frankl in received a life sentence for murder by a Cal i fornia court fo l lowing testimony by his daughter that she had recovered a memory o f witnessing h i m sexually assault and k i l l her chi ldhood friend (MacLean, 1993), a disposit ion overturned o n appeal. Nonetheless, courts have no consistent approach for assessing such claims (e.g., Porter & Marxsen, 1998; Yu i l l e & Seniuk, 1995). I n a recent Maryland case (Jane D o e et al. v. Joseph Maskel l , 1996), a court asserted that repression is not a val id phenomenon. In contrast, an Ontar io court judge has expressed m u c h faith i n the concept. Cr imina l charges were founded on allegations by a fifteen-year-old girl that her grandfather had sexually abused her between the ages o f five and eight years. T h e gir l had originally been receiving counselling f rom an assistant The Architecture of Childhood Memories 70 pastor to deal w i t h her parents' divorce during w h i c h the pastor had repeatedly asked her i f her grandfather had ever touched her inappropriately. A l t h o u g h initially the girl could recall nothing, she was informed that she had a lot o f feelings buried inside and that she should write d o w n any memories or nightmares o f the trauma. The pastor also provided her mother w i t h a 25-item list o f symptoms o f incest victims and told her that her daughter may have been abused by her grandfather. T h e v i v i d memories finally began to " f l o o d into her m i n d ' ' w h e n a man w h o looked like her grandfather walked toward her. A l t h o u g h the defendant denied the allegations, the trial judge surmised that he "h imse l f may have repressed or dissociated f rom any recollection o f what to his moral background and makeup w o u l d be repulsive and horrible acts.. . I could take it that [defendant] could be testifying honesdy as to what he recalls, and he does not and cannot recall these acts." This unsupported conjecture by the judge provided grounds for appeal and the Ontar io Appellate Cour t reversed the convict ion and ordered a new trial (Regina vs. Campbel l , 1996). I n M i s s o u r i i n 1992, a church counsellor helped Beth Rutherford " recover " a memory that her father had regularly raped her between the ages o f seven and 14 years and that her mother had helped ho ld her d o w n (Loftus, 1997b). Further, she recalled being impregnated twice and being forced to abort the fetus w i t h a coat hanger. However , later medical examination revealed that she was still a virgin at 22 and that she had never been pregnant. T h e daughter sued the counsellor and received a one mi l l ion dollar settiement i n 1996. Thi s case represented a clear example o f a created memory w h i c h had been confused for a " recovered" memory. The Canadian legal system is currendy grappling w i t h h o w to deal w i t h repressed memory cases. Estimates o f the number o f people serving prison terms based on repessed memory evidence i n Canada range f rom a handful to three dozen, although convictions have The Architecture of Childhood Memories 71 been rare i n recent years w i t h growing skepticism over recovered memories. In one Canadian case, D o n n a Cole (pseudonym) recovered memories o f being sexually abused by her father at the age o f 18 months ("False memory's v ic t ims , " T h e G l o b e and M a i l . 09 M a y 1998). H e r memories included her father placing she and her brother i n a roasting pan i n the oven, butchering and burying a female hitchiker, and anally raping the family dog after slitting its belly. T h e next meeting o f the attorneys-general i n September, 1998 is expected to address the issue o f people w h o have been convicted based on recovered memory evidence ( " M c L e n n a n to l o o k " , T h e G l o b e and M a i l . 20 M a y 1998, p. A 3 ) w i t h the possibility o f a full inquiry being ordered. There have been no studies that demonstrate the recovery o f repressed events f rom chi ldhood (e.g., Ho lmes , 1991; Paris, 1996a; Pendergrast, 1997; Pope & H u d s o n , 1995), although a few studies are commonly cited as evidence for recovered memories. Briere and Conte (1993) reported a study i n w h i c h 450 adult cl inical subjects related their sexual abuse histories. Fifty-nine percent o f respondents identified some period i n their lives before the age o f eighteen w h e n they had no conscious recall o f sexual abuse they had suffered. Melchert (1996) questioned 553 university students and found that similar proportions o f those w i t h histories o f physical, emotional, and sexual abuse reported that they had experienced periods without memory for the abuse (21%, 18%, and 18% respectively). T h e ground truth prob lem is evident i n this type o f research; i t is difficult to evaluate the accuracy o f the claims or, assuming accuracy, whether the lack o f memory resulted f rom repression or some other f o r m o f forgetting (i.e., ordinary forgetting; "suppression"). In the best-known study cited as evidence for repression, Wil l iams (1994) examined the memories o f children k n o w n to have been abused twenty years earlier. W h e n interviewed as adults about their sexual abuse histories, 38% o f the w o m e n failed to ment ion episodes o f sexual abuse w h i c h could be consistent w i t h repression The Architecture of Childhood Memories 72 (also see W i d o m , 1997). However , there are alternative explanations (e.g., Loftus , Garry, & Feldman, 1994; Paris, 1996b; Pope & H u d s o n , 1995), including the possibility that the w o m e n simply preferred not to discuss the abuse or that another form o f forgetting had occurred. Scheflin and B r o w n (1996) reviewed 25 studies o f amnesia for ch i ldhood sexual abuse (CSA) and concluded that cases o f forgetting such abuse is a robust f inding across studies and that recovered abuse memories are no more or no less accurate than continuous memories for abuse. However , note that amnesia here means only that research participants reported not remembering the incidents and offers very litde evidence for repression, i n m y view. In reviewing the evidence for and against repression, Paris (1996a) concluded that the vast majority o f recovered memories are probably created memories, w i t h the rare case representing a genuinely repressed and recovered memory. Ho lmes (1990) reviewed laboratory evidence for repression and concluded that there was no support for the concept that could not be explained by other mechanisms. Nonetheless, Gleaves and Freyd (1997) argued that the Ho lmes paper has been misapplied hundreds o f times as the definitive nail i n the coff in o f repression. It should be noted that Holmes was only interested i n the Freudian not ion o f repression and not other forms o f forgetting such as intentional forgetting or other dissociative amnesias. However , as Holmes did, we w i l l restrict our discussion to Freudian repression for the time being. There appears to m u c h more evidence against repression than i n favour o f it. First, the theoretical bases o f repression are weak. T h e unconscious processes forcing the event into an "ante- room" and the potential recoverability o f the original memory are simply not supported. T h e difficulty i n demonstrating the validity o f the not ion does not negate it ; however, more importandy here is the incongruity o f the not ion w i t h a more defensible reconstruction perspective. A s stated earlier, there is n o evidence that any memories are stored intact but rather are reconstructed f rom a person's current context. I f a traumatic experience is The Architecture of Childhood Memories 73 "repressed" during or soon after it occurs, there is clearly no opportunity for (conscious) private or social reflection o n the event. E v e n i f an event were reflected o n briefly at a deep level at the time o f its occurrence, it is highly unlikely that the original memory could endure decades i n a " d o r m a n t " state. In fact, as he rethought seduction theory, Freud himsel f was becoming aware o f distortions i n ch i ldhood memories: " W h a t seemed to be very v i v i d memories o f real occurrences i n ch i ldhood often turn out to be contradicted by clearly established facts. T h e subjective certainty manifested by his patients i n the reality o f their seduction scenes could no longer be regarded as positive evidence" (Fancher, 1973, p. 108). The psychoanalytic school conceptualized memories i n explicidy spatial terms, making two major theoretical errors, one inherited f rom Lamarck and one f rom Ari s tode (Porter & Marxsen, 1998). T h e interpretation that memories are i n the m i n d w i t h a literal spatial location is manifested i n Freud's idea o f the pr imal scene and Jung's archetype theory. Freud thought that sexual fantasies were phylogenetic " p r i m a l phantasies," remembered experiences o f our primitive ancestors (Freud, 1922). It seems that Freud thought the expression o f these fantasies was predestined and independent o f environmental triggers (i.e., an infant somehow surviving alone o n a desert is land w o u l d later experience these images just as a chi ld raised i n a family would). Similarly, Jung submitted that there are primitive images, ideas, and symbols stored up i n the m i n d and reflected i n cultures and myths, an idea attacked by Bardett (1932) as a "tangled discussion." A l t h o u g h bo th Freud and Jung were direcdy influenced by Darwin's evolutionary theory, they both were making a Lamarckian interpretation error, reasoning w h i c h D a r w i n himself had sought to overcome (Darwin, 1859). They wrongly assumed that, over the course o f human history, the impressions left by specific experiences o f humans could be passed o n to successive generations as memories stored i n the unconscious. Freud also assumed that somehow humans developed the capacity through evolution to repress particular painful experiences and ideas generally invo lv ing The Architecture of Childhood Memories 74 sexuality i n the brain. A second error o f the psychoanalytic school i n regard to memory was physiological / developmental and concerns the phenomenon o f childhood amnesia. O u r earliest memories generally go back as far as between the third and fourth birthday (e.g., K i h l s t r o m & Harackiewicz , 1982). Psychoanalysts held that infantile amnesia (a phenomenon first identified by Freud) can be explained by defense mechanisms, particularly repression, and that all ch i ldhood memories are recoverable. It is not hard to hear the echo o f Ar i s tode i n a statement by a student o f Freud , A . A . B r i l l , f rom his Basic Principles o f Psychoanalysis (1948): A s I pointed out, children take up impressions f rom the very beginning o f their existence. A s time goes on , Locke's tabula rasa becomes more and more filled w i t h them, and, like a book, the older the individual , the more voluminous it is. These traces o f early life always remain, and because they are subjected to repression, they always come to the surface more or less incomplete (p. 67). It is n o w thought that infantile amnesia is at least partially attributable to incomplete formations o f neurological connections and networks and neuron myelinization (e.g., D i a m o n d & Doar , 1989; Fischer, 1987; H o w e & Courage, 1993). F o r the first two to three years o f life the complexity o f neurological connections increases enormously. Because o f this, i n contrast to psychoanalytic theory, most early chi ldhood events, emotional or not, are not inaccessible but instead are not readily "reconstructible." Ear ly i n life, knowledge o f the w o r l d is m i n i m a l but as a child's neural networks develop and knowledge is gained, (s)he becomes inferential, or capable o f making inferences beyond the information available. Further, creative capacity increases and elaboration o f memory narratives occur as children get older (e.g., F ivush , 1994; F ivush & Haden , 1995). F o r example, i n a study by Paris and Carter (1973) children were informed that "the b ird is inside the cage" and "the cage is under the table." Later, they were provided some new sentences and some w h i c h had been heard previously. Chi ldren as young as seven years The Architecture of Childhood Memories 75 often remembered hearing the sentence "the b i rd is under the table" making a reasonable inference f rom the informat ion presented. V e r y young children are m u c h less l ikely to remember the inferences (Liben & Posnansky, 1977). The older children had constructed a memory f rom inferences based on knowledge. W h a t Mi l l e r calls "creative memory , " (reconstructive capacity), is a by-product o f cognitive development (Miller, 1983). Immanuel K a n t (trans. 1978) had long before argued that the inability to recollect early ch i ldhood memories was because sporadic perceptions had not then become unified by concepts. H a v i n g argued against the validity o f the repression concept, there may be cases o f memories for emotional early ch i ldhood events being " lo s t " and later reconstructed accurately (Knapp & Vandecreek, 1997; Porter & Marxsen, 1998). People may successfully "suppress" unwanted thoughts i n w h i c h a conscious attempt is made to put an idea or memory out o f the m i n d (Wegner & Erber , 1992). Wegner and E r b e r (1992) examined the literature o n the suppression o f unwanted thoughts and concluded that an antecedent o f thought suppression is "emot iona l i n h i b i t i o n " i n w h i c h a negative emotional expression to pain, loss, threat, other stressor is dampened or eliminated by thinking about an alternative idea. T h e suppression o f thoughts about particular events can be successful and they fail to get remembered for lengthy periods. A c c o r d i n g to the existential reconstructivist view, this might happen i f privately or socially reflecting o n an event served no adaptive function or i f adaptive qualities were outweighed by the adaptive qualities o f not reflecting o n the event. Consider the hypothetical case i n w h i c h a ch i ld is o n one occasion forced into sexual activity by the father (when the chi ld is aware that this is unacceptable). T o silence the chi ld , the father threatens the mother's life should the chi ld ever disclose. The chi ld might choose to actively try to keep the memory out o f h i s / her m i n d because telling another person w o u l d be, f rom the child's point o f v iew, counterproductive, and privately reflecting o n the painful event painful w o u l d serve litde The Architecture of Childhood Memories 76 personal benefit. T h e accurate reconstruction o f a forgotten event may be possible given the right current mental structures - i f the effects o f the forgotten event are salient and the person's currendy-held knowledge and reasoning capacities are conducive to it. F o r example, a very young chi ld w h o experienced the horrors o f a concentration camp might forget many or all events experienced there but, as an adult, come to accurately recall the gist o f them based o n what he/ she has heard f r o m significant others and the media, any lasting physical scars sustained, and so on. In fact, there are documented examples o f people losing and later " retr iev ing" memories o f past abuse (Knapp & Vandecreek, 1997). M o r e convinc ing cases include those i n w h i c h people have recalled abuse spontaneously or through therapy and then later the perpetrator confessed (we w i l l assume that he was not experiencing a created memory) or other corroborating evidence was found (Martinez-Taboas, 1996). T h e rate o f this phenomenon is u n k n o w n (Knapp & Vandecreek, 1997), and to my knowledge there has never been a documented case o f retrieving an accurate emotional memory f rom infancy. Katherine N e l s o n (1993), a leading expert on chi ld development, discussed the developmental origins o f autobiographical memory i n relation to infantile amnesia. She argued that young children learn to share memories w i t h others as they acquire the narrative forms o f memory sharing. Such sharings or recountings are effective i n reinstating experienced memories only after children can use another person's representation o f an experience i n language as a reinstatement o f their o w n experience. R o b y n F i v u s h and colleagues (e.g., F ivush , 1994; F ivush & Haden , 1995) have found that young children's narratives about the same event become more elaborate and complex over time, especially between the ages o f 5 and 6 years. In accordance w i t h a reconstructivist perspective, sharing memories helps to create the memories themselves, something not possible i n infants. The Architecture of Childhood Memories 77 W h a t does the public think o f recovered memory reports? Research o n people's perceptions o f witnesses indicates that laypersons have traditionally given m u c h credence to accounts o f events and identification o f perpetrators (e.g., B e l l & Loftus , 1988). V i v i d accounts w i t h a high degree o f expressed confidence have been found to be especially persuasive (e.g., P e n r o d & Cuder , 1995). V e r y litde research has been conducted o n the credence given to confident accounts o f ch i ldhood events. Loftus , Weingardt, and H o f f m a n (1993) examined h o w people react to claims o f recovered chi ldhood memories. Participants answered questions about a legal case arising out o f allegations o f sexual assault. In one version o f the allegation, the complainant claimed to have repressed the memory for ten years and recovered it i n therapy whereas i n the other, she showed continuous recall for the incidents but d id not report them for ten years. Participants, particularly males, judging the allegations tended to be more skeptical o f the repression case. A s wel l , the repressed and non-repressed cases evoked different considerations; participants skeptical o f the repressed memory viewed it as honest but false, whereas skeptics o f the non-repressed memory thought that the complainant may have been lying. U s i n g a similar approach, K e y , Warren , and Ross (1996) found that participants were more likely to convict a defendant w h e n testimony concerned a continuous memory for sexual abuse (67%) than a recovered repressed memory (58%). There appears to exist an acceptance o f the concept o f repression but a m i l d skepticism for particular cases among university students (cf. Clark & Nightingale, i n press; G o l d i n g , Sego, Sanchez, & Hasemann, 1995), although i n both o f these studies a large number o f participants convicted the defendant based o n the recovered memory evidence. Some psychologists have felt sufficiendy confident i n their ability to identify recovered repressed memories that they have provided guidelines publicly. F o r example, fo l lowing the highly controversial 1995 C B C Passionate Eye documentary " D i v i d e d M e m o r i e s " (Beckell, The Architecture of Childhood Memories 78 1993) the host asked a panel o f mental health professionals the question o n all the viewers' hps: "Is there any formula...is there any way to tell w h e n a memory is real?" A research psychologist reported to viewers that there are two specific attributes almost invariably associated w i t h true recovered memories o f abuse - appropriate levels and types o f emotional and sensory experience w i l l accompany the memory being recovered and the memory recovery w i l l be involuntary ("out o f the blue") and a fragmentary nature, rather than sudden and complete. H o w e v e r , it is currendy not clear whether retrieved memories for real events are qualitatively similar to " cont inuous " memories. Presumably, continuous memories w o u l d have undergone more temporally-contingent distortion (from social factors and so on) than newly retrieved memories. E idet ic memories T h e idea that certain individuals have photographic or eidetic memory has been controversial throughout this century (see Yarmey, 1979). Eidet ic imagery is characterized by an exceptionally v i v i d k ind o f visual image w h i c h has the clarity o f perception (Yarmey, 1979), defined as " the ability, possessed by a minori ty o f people, to 'see' an image that is an exact copy o f the original sensory experience" (Kagan & Havemann, 1972, p. 588). Supposedly, the eidetiker moves h i s / her v i s ion over the image as though the scene had been stored i n its original state. E idet ic memory, although still discussed i n introductory psychology textbooks, has been largely dismissed by some memory theorists (e.g., Leask, Haber , & Haber , 1969; Neisser, 1967) because perfect recall is rarely, i f ever, found i n supposed eidetikers despite earlier claims o f veridicality (e.g., K l i i v e r , 1926; Yarmey, 1979). Thus , this not ion , like " f l a shbulb" memories, must be met w i t h skepticism and holds litde utility i n any consideration o f h o w people recall emotional events. There is certainly no evidence that v i v i d memories for The Architecture of Childhood Memories 79 chi ldhood events can represent an exact copy o f the original experience, although some people undoubtedly recall ch i ldhood events better than others. T h e strange case o f S. A classic case w h i c h seems to offer support for the spatial v iew was presented by the Russian psychologist, Aleksandr Lur i a (1968) - the case o f S. f rom the 1930s. S. appeared to have a limidess memory akin to a tape-recorder, lacking schemata and retaining al l i n c o m i n g informat ion, regardless o f the personal relevance o f the information. Perfect retention allegedly occurred for everything f rom a string o f nonsense syllables to the words i n a personal letter and the capacity o f his memory allegedly had no direct limits. F o r example, a long list o f words w o u l d be recalled accurately months or years after the initial presentation w i t h no warning prior to the test. S. apparendy was able to convert sounds into v i v i d visual imagery, a process Lur i a called "synesthesia." A s soon as he heard or read a w o r d , it was immediately converted into a visual image corresponding w i t h the object the w o r d signified for h i m . In accordance w i t h Aristotel ian theory, he had no need to " m e m o r i z e " but only had to "register an impress ion" w h i c h he w o u l d then " r e a d " on a later date, like a super-eidetiker. A c c o r d i n g to L u r i a , S. d id not experience the normal amnesia for the first few years o f life but instead had v i v i d recollections dating to infancy. Does S. provide evidence for a storehouse view o f memory? A l t h o u g h still a very poorly understood case, there are reasons to believe that, at least i n terms o f his autobiographical memories, S. was exceptionally proficient at reconstructing rather than simply reading the original sensory perception. A l t h o u g h L u r i a trusted the chi ldhood memories o f S., his documented accounts suggest aspects o f reconstruction (record o f August , 1934, p. 77): The Architecture of Childhood Memories 80 I was very young then...not even a year o ld perhaps...What comes to m i n d most clearly is the furniture i n the r o o m , not all o f it, I can't remember that, but the corner o f the r o o m where m y mother's bed and m y cradle were. A cradle is a small bed w i t h bars o n bo th sides, has curved wickerwork o n the under part, and i t rocks...I remember that the wal l-paper i n m y r o o m was b r o w n and the bed white...I can see my mother taking me i n her arms, then she puts me d o w n again...I sense movement.. .a feeling o f warmth, then an unpleasant sensation o f cold . . . / see myself in m y mother's bed, first w i t h m y head toward the wal l , then facing the door...I recognize the sound my o w n voice makes. O n e clue to reconstruction is perspective change, that S. can, like many people, see himself i n the v i v i d images o f many o f his ch i ldhood memories (e.g., R o b i n s o n &c Swanson, 1993). O f course, i f he were simply "reading the impressions," his perspective w o u l d disallow seeing himsel f i n the image. Nonetheless, S. represents an enigma f rom any perspective o n memory. O n e o f the ideas o f the present paper is that the memory image is an architectural product i n w h i c h verbal and other information are transformed into a sensory image reconstructed for retrieval, a process w h i c h seemed, i n the case o f S., to be impeccable and unmediated by schemata. T o summarize, flashbulb memories, repressed memories, and eidetic memories are still c o m m o n l y given credence i n psychology and, especially repressed memories, brought up i n the courtroom. Conceptually, all three are variants o f the storehouse not ion , have not received empirical validation, and are difficult to defend scientifically. Flashbulb memories do not seem to represent as unique class o f emotional memories and can be far f rom exact reinstatements o f the original event. The repressed memory as formulated by Freud and his predecessors, although virtually unprovable, is an unlikely phenomenon and many cases o f supposed The Architecture of Childhood Memories 81 recovered memories have turned out to be created memories (e.g., Loftus , 1993). A n alternative explanation to repression is that intentional suppression o f thinking an event, although sometimes this process can have the opposite effect (Wegner & Erber , 1992). A forgotten event could theoretically become accurately reconstructed based o n current knowledge structures, in the words o f Husserl . T h e existence o f eidetic memories has not been proven w i t h investigations indicating many errors i n "eidetic" memory reports. C h i l d h o o d memories for emotional events may retain a v i v i d quality but this is not wel l explained by any o f the above conceptions. In sum, emotional chi ldhood memories are better explained as a result o f the reconstructive processes discussed earlier. The Architecture of Childhood Memories CHAPTER FOUR The Architecture of Emotional Memories The Architecture of Childhood Memories 83 " T h e horror o f that moment , " the K i n g went on, " I shall never, never forget!" " Y o u w i l l , though, " the Queen said, " i f you don't make a memorandum o f i t . " (Alice i n Wonder land . Lewis Carrol) Basic Factors Af fect ing M e m o r y Quality H a v i n g set out some basic principles o f memory reconstruction and expressing skepticism o n dominant spatial theories, differences between real and created memories for emotional events w i l l be further considered. M a n y researchers have attempted to identify w h i c h characteristics o f ch i ldhood events lead to enduring or v i v i d memories. T h e importance o f w h e n and h o w often an event is reflected u p o n has been alluded to i n the earlier discussion. That is, there is a relationship between w h e n events are reflected u p o n and the frequency o f reflection. A s Marce l Proust (1981) observed, "we soon forget what we have not deeply thought about." Other factors w h i c h seem to influence h o w people construct emotional memories are next described, inc luding novelty, affective potence and valence, imagery formation, the memory trace, and time passage. The differential reconstruction o f emotional events by adults and children is then explored, and some specific features o f extreme emotional memories (trauma) are examined. N o v e l t y Conway and Bekerian (1988) reported that personally important events were recalled more v iv id ly than personally unimportant events, i n line w i t h the reconstructivist perspective. In B r o w n and Ku l ik ' s (1977) paper on flashbulb memories, it was argued more specifically that v i v i d memories are created for events w i t h novelty, consequentiality, and emotional arousal. A s w i l l be demonstrated though, an event's novelty does not render it immune f rom distortion. Nonetheless, the novelty o f any event is an agreed-upon determinant o f its subjective The Architecture of Childhood Memories 84 memorability. A s one might expect, l o w frequency events are generally recalled more accurately than high-frequency events (e.g., Brewer, 1994; L i n t o n , 1979; Wagenaar, 1986; W h i t e , 1982). In one study o f randomly-sampled autobiographical memories, positive associations between memory accuracy and low-frequency events and low-frequency locations were found (Brewer, 1988). Unusual , bizarre, or highly distinctive events appear to be better reconstructed than mundane, ordinary events (e.g., Schmidt, 1991), sometimes referred to as the von Restorff effect (1933). A l o n e , the observation that novelty is associated wi th enduring memories is not that informative. Because nove l events can range f rom innocuous (visiting N e w f o u n d l a n d for the first time) to harmful (being i n a car accident), it is, o f course, important to consider emotional factors i n the experience but note that there is undoubtedly a significant correlation between the novel and emotional aspects o f autobiographical memories. N o v e l events may often have an aspect o f personal uncertainty and may be reflected on frequendy i n a relatively lengthy appraisal process. Inferential distortions associated specifically w i t h schemata, w i l l probably not occur i n reconstructions for nove l events (unless the person has acquired vicarious informat ion o n the event category). Affect ive strength and valence Important factors i n the creation o f enduring autobiographical memories appear to be level and type o f emot ion associated w i t h the events (see Christianson, 1992; W i n o g r a d & Neisser, 1992, for reviews). It is clear that highly emotional events are associated w i t h superior reconstruction ability (e.g., Wagenaar, 1986; W h i t e , 1982; Yu i l l e & Cutshall , 1986), although there a number o f different memory patterns for emotional events (Yuille & Day len , 1998). A s mentioned, Yu i l l e and colleagues (Yuille & Cutshall , 1989; Yu i l l e et al., 1994) have discussed remarkable memories characterized by endurance over time, likely resulting f rom repeated The Architecture of Childhood Memories 85 reconstructions o f the event, and recalling the event soon after it occurs (Yuil le & Tollestrup, 1992). E m o t i o n a l valence (good versus bad) also appears to influence the reconstruction o f life events. E m o t i o n a l arousal sternming f r o m negatively-valenced, personally important experiences has been shown to create enduring, detailed memories relative to memories for neutral or positive personal experiences (Enright, 1989; Kos s , Figuerdo, B e l l , Tharan, T r o m p , 1996), w h i c h perhaps makes good evolutionary sense. Enr ight (1989) categorized participants' reported memories into thematic clusters and found that negative emotional life events such as accidents, injuries, illnesses, and operations produced more subjectively v i v i d memories than positive events such as a vacation. O n the other hand, Sheingold and Tenney (1982), i n one o f the few long-term studies o f memory for salient chi ldhood events, found that no loss o f informat ion had occurred i n four-year-old children's memories for the bir th o f a sibling over a sixteen-year period. O f course, we can only presume that the birth o f a sibling is appraised by four-year-old children as a positive personal event (the possibility o f sibling rivalry could also be considered). Overa l l , though, the literature appears to indicate a superiority o f memory for negative life events. In K i h l s t r o m and Harackiewicz ' (1982) survey o f h igh school and college students regarding their earliest personal recollection, there were proportionately more memories classified as " traumatic" than other categories among the high school students, indicating the endurance o f negative emotional memories. Enr ight (1989) attributed the difference to a specially-evolved mechanism that allows humans to accurately reconstruct hfe-tlireatening events. She observed: "whereas look ing at the G r a n d C a n y o n may be a distinctive and memorable event i n one's life, it is not generally associated w i t h the affective and cognitively distinct experience o f thinking, " * @ # & , I'm going to die!!" Positive events such as a wedding have existential relevance but not to the extent o f dangerous experiences such as getting lost or being v ic t imized through violence. D u r i n g our evolutionary development, survival may have The Architecture of Childhood Memories 86 been enhanced by recalling negative circumstances to avoid repeating them and creating strategies for escaping them i n the future. A l t h o u g h emotional arousal appears to be a causal factor i n the creation o f v i v i d memories, the relationship may not be direct. It is unlikely that the phenomenology o f v i v i d memories is entirely dependent o n affect given the results o f a simple litmus test negating it - v iv id memories i n people w h o experience litde emotional experience. A c c o r d i n g to clinicians and researchers, a keystone o f psychopaths is their marked absence o f normal human emotional experience (e.g., Cleckley, 1982; Hare , 1993, 1996; Har t & Hare , 1997; Porter, 1996). D o psychopaths ever reconstruct autobiographical events vividly? I f so, strength o f emotional responding during the event cannot be the sole causal factor i n memory vividness. T h e answer is a macabre yes. Some psychopaths reconstruct their crimes vividly and accurately (Porter, Yu i l l e , B ir t , & Herve , submitted for publication). Psychopaths Cl i f ford Ol sen and H e n r y Lee Lucas recalled many o f the murders they committed accurately years later. Lucas, convicted o f twelve homicides i n 1987, reconstructed many o f crimes unerringly years after their occurrence. H e provided incredibly detailed reports w h i c h often matched physical and witness evidence precisely, and investigators were "amazed at his perfect reca l l " (Cox, 1991). T h e crimes may have been personally meaningful to these psychopaths; they likely rehearsed the crimes before they had occurred during fantasy engagement and subsequendy reflected u p o n them frequendy (Porter et al., submitted for publication), facilitating reconstruction. Af fec t and existential appraisal O n e idea that has received litde attention is that emotional arousal is not direcdy related to quality o f memory but rather is a gauge o f the pertinence a particular event has to an mdividual's well-being. That is, a strong emotional reaction (fear, excitement) might serve to The Architecture of Childhood Memories 87 facilitate init ial appraisal and increase a person's awareness that this event has high personal value, triggering a process o f reflection to self and others. T h e endurance o f emotional memories could be related more to this depth and frequency o f reflection than the emotional state itself. That is, i f a person was really frightened w h e n lost i n the woods , (s)he is more likely to recall the experience privately and socially, fulfill ing the communicative function (avoidance o f that dangerous circumstance). Such memories may endure i n terms o f vividness although memory for some details may become become distorted. O n the other hand, there is considerable evidence that emotion and m o o d have specific effects o n reconstruction capacity. F o r example, affect can represent the "state" i n state- or mood-dependent memory (e.g., Bower , 1981,1992; E i c h , 1989; E i c h & Macaulay, 1994). In 1996, the present author along w i t h a forensic psychologist (Dr. J o h n C . Yuil le) interviewed an elderly male w h o had murdered his wife (Porter et al., submitted for publication). H i s involvement was not challenged, as he readily admitted it. A l t h o u g h he had no cr iminal record, twenty years earlier police had been called to his home fo l lowing a report o f assault by his wife. H e reported to us that o n the day o f the murder, his wife had complained bitterly about a b i l l received i n the mai l . A n ensuing argument led to h i m exiting the home and breaking a botde o n a tree. The wife then stated her intent ion o f p h o n i n g the police and he re-experienced the rage, fear, and humil iat ion he had experienced the first time his wife had phoned them. A t this point, the suspect's memory began to fade. H e recalled a shoving match fol lowing his demand that she put the phone d o w n . She proceeded to spit i n his face after w h i c h he is unable to reconstruct events unt i l waking up next to her lifeless body. It is quite possible that the man's inability to recall the murder was related to the affective state he experienced at the time. In Proust's nove l Remembrance o f things past, there is a we l l -known scene i n w h i c h , late i n life, the narrator is sitting wi th his mother while enjoying tea and cake. Suddenly, his m i n d is The Architecture of Childhood Memories 88 flooded w i t h v i v i d images o f his chi ldhood. W h i l e he struggles to understand this strange experience, he realizes that the w a r m tea and cake crumbs tastes the same as petite madeleines, the litde cakes his aunt used to give h i m after dipping them i n her cup o f tisane. This taste had opened up what seemed to be a treasure-chest o f ch i ldhood memories for the narrator. State-dependent memory ( S D M ) is a well-documented phenomenon (see E i c h , 1989; Bower , 1992, for reviews). Information acquired i n a particular m o o d state is more easily reconstructed i n a similar affective state than a different one (Bower, 1981). B o w e r (1992) reviewed the literature and came to several conclusions about S D M : (1) the moods induced must be intense and fairly different f r o m one another for the occurrence o f S D M ; (2) S D M is most pronounced i n people recalling autobiographical events and less so i n artificial laboratory experiments; and (3) S D M is rarely seen w i t h recognition tests or strongly cued recall tests but is most likely w i t h free recall. F r o m a reconstructivist view, a similar affective state is conducive to reconstructing the original circumstances w h i c h elicited the response, a rational knowledge-based process i n w h i c h the original perception is not called forth but rather re-created according to varying conditions. ( A n interesting assertion o f Schopenhauer's [1851/ 1951] reversed the process; it is the emotions that are reconstructed f rom knowledge o f what has transpired during an autobiographical event). T h e images o f memory T h e light o f memory or rather the light that memory lends to things, is the palest light o f all...I am not sure whether I am dreaming or remembering , whether I have l ived m y life or dreamed it. Just as dreams do, memory makes me profoundly aware o f the unreality, the evanescence o f the wor ld , a fleeting image i n the m o v i n g water. - Ionesco (1968) T h e ability to transform verbal ideas and other semantic informat ion into visual and other sensory images is an important feature o f memory. Recal l the interesting case o f S. w h o was apparendy an expert at synesthesia, to use Luria's term. Immediately upon encoding a The Architecture of Childhood Memories 89 stimulus such a w o r d , a corresponding visual image appeared i n S.'s m i n d . T h e study o f visual imagery has been a major focus wi th in cognitive science for decades. A l t h o u g h pr ior to the 1960s, imagery was studied primarily through introspective reports, chronometric and physiological methods have since been developed (e.g., Murray & Ba ldwin , 1996). In some studies, participants have been required to construct a visual image after receiving verbal informat ion (e.g., Kos s lyn , Cave, Provost , & v o n Gierke , 1988). React ion time and active brain areas have been gauged, w i t h results indicating that visual images are created sequentially rather than concurrently and i n areas o f the brain involved i n visual perception (Murray & Ba ldwin , 1996). F o r example, i n a recent study, participants were asked to construct drawings o f geometric figures after having seen such figures or heard them described. Ful ly accurate representations were reconstructed i n bo th conditions (suggesting the possibility o f similar mental images being generated) but the product ion o f the drawings took longer i n the verbal condi t ion (Murray & Ba ldwin , 1996). However , such figures are m u c h less complex than autobiographical events and the authors d id not compare the image depictions thoroughly o n qualitative features. Zaparniuk et al. (1995) uti l ized a more complex stimulus. Forty subjects either viewed a videotaped event and recalled i t or listened to the tape and recalled i t as though they had viewed it. T h e reconstructions differed qualitatively i n a number o f respects. F o r example, the group w h i c h had v iewed the videotape recalled i t i n a more spontaneous fashion without a r igid structure. T h e idea that the perception o f an object or event leaves a lasting " m a r k " i n the brain is, o f course, an o l d one such as the wax tablet metaphor used by the ancient Greeks. M o d e r n psychologists have long grappled w i t h a theoretical construct called the memory trace, described i n M a l c o l m (1977): .. .a person witnessed an event or underwent some experience at a certain time. Le t us express this by saying that an event, E , occurred at time t\. A t a later time this person The Architecture of Childhood Memories 90 either relates event E , or i n some other way, manifests a memory o f E . Le t us call this second event "the memory response," M , w h i c h occurs at time t3. T h e temporal interval between t i and t3 might be an hour, a day, a m o n t h , or a year. A trace theorist assumes that the event E produced a memory trace o f E (call it T) w h i c h persisted i n Person, P , f rom ti to t3. Nowadays it is typically assumed that T is a state or process o f the central nervous system. The occurrence o f some stimulus, S, "excites" or "activates" the trace, T , and this causes the occurrence o f M (p. 171). Current trace theorists h o l d that the trace is not only something persisting after the original event occurs but is something that is representative o f the past event, usually conceived o f as a neural representation. This trace has a causal role i n eliciting a memory, provides for retention o f informat ion about events, and possesses a similarity i n structure to the original perception (Malco lm, 1977), like the grooves o n a phonograph. In forensic psychology, as earlier noted, there is an ongoing debate about the effect o f mis information o n memory. O n e camp believes that post-event misinformation replaces or overwrites the original informat ion whereas the other argues for the existence o f source confusion (e.g., Johnson , Hashtroudi , & Lindsay, 1993). Th i s latter v iew is more l ike a trace theory (Hyman & Pendand, 1996) i n w h i c h the traces o f both the original and new information are stored independentiy i n the brain. W h e n attempting to retrieve the original accurate information, the mis information may come to m i n d and be considered, erroneously, to be the original. T h e problem o f source moni tor ing was highlighted i n C r o m b a g and Wagenaar's (1996) recent study w h i c h showed that a high proport ion o f participants recalled witnessing a jet crash o n television when , i n reality, they had only heard about it f rom others. I n A k i r a Kursawa's f i lm Rashomon. reflecting a not u n c o m m o n phenomenon i n forensic arenas, four eyewitnesses viewed the same event but each person's recollection differed The Architecture of Childhood Memories 91 fundamentally f rom the others, despite each expressing complete confidence i n the memory. O n e w o u l d predict, i f the traces for the original event were at all similar and i n effect, the memories w o u l d also be similar to some degree. In a real-life case, a high-profile 1995 investigation wi th in the Canadian navy looked at allegations that a high-ranking officer had inserted a cigar-tube into the rectum o f an unconscious shipmate during a party. In the investigation to fol low, witnesses present at the time o f the act i n question gave extremely varying accounts o f h o w the incident had transpired, f rom the physical pos i t ion o f the v i c t i m to what was actually done w i t h the cigar tube, although many expressed that they had experienced shock and embarrassment at the time o f the memorable event. Presumably, the perception o f the original event was similar for many o f the witnesses, leading to similar "traces." H o w then could all their memories have been drastically different i f they had been based o n similar original traces for the event? It appears that there was more than peripheral reconstruction or distortion. T h e mental structures o f the witnesses were likely very different, especially given the seriousness and potential ramifications o f the allegation, reflected i n their memoria l imagery. Consider ing complete created memories, there is presumably no trace whatsoever. O n the other hand, it is possible that created memories are constructed f rom memories o f other real events (Loftus, 1997a, 1997b). T h e resolution o f the trace issue is still far off. A t this point , the no t ion o f the trace seems to have l imited utility i n facilitating our understanding o f people's recollections for emotional events. Eventually, the trace idea may be reconciled w i t h the incontrovertible reconstructive nature o f memory. The Architecture of Childhood Memories 92 Passage o f time and perspective change T h e architecture o f autobiographical memories is a function o f time, as noted by Brewer (1996): ...recent personal memories retain a relatively large amount o f specific informat ion f rom the original phenomenal experience (e.g., location, point o f view) but that w i t h time, or under strong schema-based processes, the original experience can be reconstructed to produce a new non-veridical personal memory that retains most o f the phenomenal characteristics o f other personal memories (p. 44). G i v e n that the product o f reconstruction evolves over time, emotional memories for ch i ldhood events recalled i n adulthood should be subject to more distortion than emotional events experienced i n adulthood. O v e r the years, distortion w i l l result f rom retroactive influences i n the absence o f well-established schemata and cognitive complexity i n children, discussed below. Di s tor t ion occurring o n memories created during adulthood w i l l be more contingent o n the proactive influences o f prior knowledge and pre-existing memories. Events o f an emotional nature, having the greatest personal relevance, w i l l be reflected on often f rom ch i ldhood into adulthood and w i l l be subject to considerable evolution. Perspective change is a time-related phenomenon w h i c h has direct relevance i n understanding reconstructive processes and potential relevance i n memory assessment. That is, i n many ch i ldhood memories people can see themselves i n the visual image (Nigro & Neisser, 1983; R o b i n s o n & Swanson, 1993), possibly first described by Freud i n his observation o f "screen memories . " A c c o r d i n g to Freud (1974a), self-protecting screen memories have four principle features: (1) they lack feeling tone; (2) repetitive recall; (3) they are predominandy visual; and (4) the person sees h i m / herself i n the visual image. This represents an "observer" perspective as opposed to a "part icipant" perspective, i n w h i c h the scene appears f rom one's The Architecture of Childhood Memories 93 o w n posi t ion " l o o k i n g outward. " A n observer perspective offers prima facie evidence that a degree o f distortion has occurred, and represents some o f the best evidence for an existential reconstructive perspective. T h e first empirical studies investigating observer vs. participant perspectives i n autobiographical memories were reported by N i g r o and Neisser (1983). In their pi lot study, they surveyed people about w h i c h perspective characterized their memories for ten categories o f events. T h e field (observer) perspective was more c o m m o n overall; the situation produc ing the most participant memories was having a conversation w i t h a friend (65%), whereas the situation generating the most observer memories was giving a public presentation (55%). In another study (Study 3), observer memories were more likely for such circumstances as, "wa lk ing or running f r o m a threatening situation", "g iv ing an individual public presentation", whereas field memories were more likely for events such as, "demonstrating a skilled act", "watching the news" , and "watching a horror m o v i e . " Events were more likely to be recalled f rom an observer perspective i f they had occurred less recendy, indicating that memory images evolve over time. . In a second study exarriining perspective i n autobiographical memories, R o b i n s o n and Swanson (1993) were interested i n whether people could intentionally alter perspective. They found that changing was easier for recent or vividly recalled events. T h e observer perspective was reported for 59% o f ch i ldhood memories. 47% o f intermediate life memories, and only 26% o f recent adulthood memories. Secondly, the affect experienced was altered w h e n perspective intentionally changed, w h i c h all participants were able to do. Reported affective experience decreased w h e n an observer perspective was adopted. This l imited but important research re-affirms that distortion is temporal, that events are reconstructed, and that the meaning attached to events alters over the lifespan. Di s tor t ion can be vol i t ional i n that perspective changes can consciously and intentionally occur. The implications o f these findings The Architecture of Childhood Memories 94 for memory assessment are considerable and the observer/ participant d ichotomy is a dependent measure to be explored i n the present study. W h a t could be predicted about differential perspectives i n memories for real, fabricated, and false ch i ldhood events wi th in individuals? This has never yet been addressed so we are proceeding o n a speculative level. It could be predicted that observer memories w i l l occur more often i n frequendy-reflected emotional events, w h i c h generally contain an element o f danger and w h i c h occurred long ago. It is possible that, compared to fabricated and false events, true emotional events o f ch i ldhood w i l l be more likely to be recalled f rom an observer perspective given the length o f time past. A l s o , a reasonable predict ion f rom the existential v iewpoint is that emotional events represent an element o f "danger" and viewing one's behavior i n the reconstructed image may be adaptive. In lied-about events, n o such danger w o u l d have been present and self-assessment through the observer perspective w o u l d have litde adaptive value. I f a person has seen photographs o f h im/her se l f i n a particular chi ldhood situation, (s)he may be more likely to v iew h im/her se l f i n the memory. U n l i k e fabrications, unintentional memories for false ch i ldhood events created i n adulthood may be more likely to be characterized by the observer perspective given that their source is often external. F o r example, i f a family member falsely informed me that they recall seeing me o n m y father's shoulders r iding o n a black horse, I might very wel l recall the scene that way. Thus , a tentative prediction for this criterion is that wi th in individuals, the observer perspective w i l l be least frequent i n memories for fabricated ch i ldhood events and most frequent for false ch i ldhood events, wi th memories for real events falling somewhere i n between (possibly approximating the 59% reported by Rob inson & Swanson, 1993). The Architecture of Childhood Memories 95 D r e a m i n g and other altered states T h e occurrence o f reconstructive processes may not be restricted to the conscious state. Recal l Husserl's example o f a detective, stumped over the murder case i n the evening, w i t h his epiphany happening during sleep. Sleep and its meaning have long evaded science, although i n 1900, F reud (1974b) proposed an elaborate theory o f dreams as censored outiets for unconscious desires i n T h e interpretation o f dreams, whose tenets are still accepted by many. M o r e recent conceptions have been that dreams represent the brain's "scavengers," by w h i c h unnecessary informat ion is eradicated or that they result f rom random neuronal firing. Th i s mtriguing latter conception could be interpreted as dreaming being the well-adjusted person's psychosis or that the psychotic person is essentially i n a "sleep state" while awake. W i n s o n (1997) proposed that dreaming acts as "the means by w h i c h animals f o r m strategies for survival and evaluate current experience i n light o f those strategies" (p. 58). In other words , sleep (during R E M ) is generally an uninterrupted time o f reflection, integration o f experiences, and conclusions for opt imal strategies to cope w i t h threats to existence. A person can wake up w i t h different schemata/current mental structures than before falling asleep. Other altered states such as a lcohol and marijuana intoxicat ion also influence h o w people perceive and reconstruct experiences (e.g., Y u i l l e , Tollestrup, Marxsen, Porter, & Herve , 1998). Remember ing a real event appears to be architectural, the perceptual informat ion serving as the blueprint or scheme. In the hands o f reasoning and knowledge, an image is progressively constructed unt i l the final work , a memory, is created. T h e memory may change w i t h each recreation. N e x t , reconstruction factors specific to adult and chi ld memories are considered w i t h predictions o f the features o f memories for adults reconstructing ch i ldhood emotional events. The Architecture of Childhood Memories 96 T h e Reconstruction o f E m o t i o n a l Events B y Adult s F e w previous studies have looked specifically at h o w adults reconstruct stressful events f rom ch i ldhood but many have examined adults' and children's recall for emotional events o f adulthood and chi ldhood, respectively. A s mentioned earlier, some research indicates that negative events are more enduring i n memory than positive events (see also K o s s et al., 1996). Unfortunately, the effect o f stress o n memory is unresolved, pardy because remaining ecological validity concerns. F e w studies have been conducted o n the effects o f extreme stress o n memory because o f the obvious practical and ethical preclusions involved. Part o f the p rob lem is the definition o f "stress." I n many studies l o o k i n g at adults, stress was induced by the application o f white noise w h i c h is independent o f the event to be recalled (see Y u i l l e et al., 1995). Other researchers have used violent depictions wi th in films to test the effects o f stress o n eyewitness memory (e.g., Loftus & Burns, 1982). G i v e n that most people today are inured to witnessing such violence, the generalizability o f this approach is questionable. F o r example, Myers (1990) reported that by high school graduation, the average A m e r i c a n witnesses about 25000 violent deaths o n television and i n films. Meta-analyses have yielded equivocal findings. F o r example, Deffenbacher (1983) found that ten o f 21 studies he reviewed suggested that higher levels o f arousal enhanced or left unaffected eyewitness accuracy whereas eleven studies indicated that heightened arousal impaired accuracy o f recall. Notab ly , the majority o f studies i n w h i c h memory was enhanced by stress had employed a l ive witnessed event whereas only one o f the eleven studies showing a debilitating effect had employed a live event. M o r e realistic field studies generally indicate that personal emotional events tend to be recalled more accurately than ordinary events. The Architecture of Childhood Memories 97 O n e robust f inding o n emotional stress and memory is that stress tends to focus perceptual attention (e.g., Easterbrook, 1959; Loftus , 1980; Yui l l e et a l , 1995; Y u i l l e & Daylen , 1998). That is, as a person becomes stressed, attention tends to narrow o n either the source o f the stress or its effects, rendering a "stressor" independent o f the event more o f a distractor (Yuille et al. , 1995). Recent studies o f people w h o have experienced high levels o f stress i n v ict imizat ion contexts (e.g., victims o f armed robbery, see Tollestrup et al., 1994) suggests that the focussing impact o f emotional stress is complex. Some victims o f a stressful event may ho ld v i v i d , detailed memories for the events, whereas others may only recall their subjective emotions experienced during the event rather than external details (Yuille & Daylen , 1998; Y u i l l e & Tollestrup, 1992). There is also some evidence that memory for information associated w i t h emotional events, such as details preceding and succeeding the events or peripheral informat ion wi th in the event, is retained less accurately than the critical, central aspects o f the event w h i c h may endure the test o f time (Christianson, 1992; Christianson & Ni l s son , 1984, 1989; Yu i l l e & Tollestrup, 1992). A person reconstructing an event o f w h i c h only central details were originally perceived is likely to chiefly draw u p o n the knowledge o f this information. D i s tor t ion is expected to occur over time as the meaning o f the event changes for the person but the central informat ion may be reconstructed accurately. A pioneering field study was conducted by Yu i l l e and Cutshall i n 1986 (Cutshall & Yui l l e , 1989; Y u i l l e & Cutshall , 1986) i n w h i c h the memories o f thirteen eyewitnesses to a homicide were assessed four to five months after the crime. B y comparing the accounts to forensic evidence and police reports, they found that the witnesses gave accurate, detailed reconstructions o f the event (accuracy rates o f 83%, 76%, and 90% for action details, person descriptions, and object descriptions, respectively) and were highly resistant to misleading questions. A d o p t i n g a similar approach, Christianson and Hubinette (1993) examined witnesses' The Architecture of Childhood Memories 98 (victims' and bystanders') ability to recall post office robberies. Results indicated that recollections o f details (actions, weapon, clothing) four to fifteen months after the crime were highly consistent w i t h the initial reports to police. These studies indicated that heightened emotion can sometimes have facilitating effects o n memory contrary to claims o f general deteriorating effects (e.g., C l i f ford & H o l l i n , 1981; C l i f ford & Scott, 1978; Siegel & Loftus , 1978). These results, i n w h i c h people report o n highly threatening experiences, are consistent w i t h the existential reconstructivist view. It w o u l d be expected that the witnesses reflected o n the event privately and to others (some may have communicated w i t h fel low witnesses) often, serving the communicat ive function (e.g., "here's what the guy did , here was my reaction, here was the outcome, and here is what I may have done differendy"). A study by Wagenaar and Groenewed (1990) yielded findings consistent w i t h the idea that stress can facilitate event reconstruction. Tak ing advantage o f a remarkable research opportunity to study memory for highly stressful events, they examined the memories o f 78 concentration camp survivors for their experiences i n the Second W o r l d W a r i n a case against Marinus D e Rijke. The i r memories were found to be accurate and detailed, or "generally well-remembered" (p. 77), i n 1984-87 w h e n compared to their reports to Nuremberg investigators during and soon after the war (1943-47). Account s o f the conditions i n the camp, malicious treatment, daily routine, labour, housing, and main guards were "remarkably consistent" (p. 84) over four decades. Nonetheless, the authors chose to emphasize the "essential details" that were often forgotten such as forgetting names and dates to conclude that "apparent intensity o f experiences is not a sufficient safeguard against forgetting" (p. 77). In my view, it is impressive that 16/30 witnesses produced the correct camp registration number when asked to do so. This pattern o f findings actually supports the idea that memory for extremely stressful (or traumatic) events is enduring and accurate. The Architecture of Childhood Memories 99 A study appeared recendy suggesting that reconstructions for highly emotion-laden events such as rape and other negative experiences were less detailed, v i v i d , and coherent than memories for positively-valenced events. K o s s et al. (1996) examined characteristics o f memories formed i n response to rape and other intense unpleasant experiences and memories for pleasant events. A large sample o f participants were forwarded a questionnaire asking whether they had ever experienced an attempted or completed rape. I f such an incident had occurred, participants were asked to recall it and provide their cognitive appraisal o f the incident. I f not, participants were asked to choose another intense life experience to recall. A l l respondents then described their memory for the experience o n criteria based o n the reality moni tor ing m o d e l o f Marc i a Johnson and colleagues (e.g., Johnson, 1988). Results f r o m the first sample and conf i rmed using a cross-validation approach, indicated that the memories were distinctive i n many ways. T h e rape memories, and to a lesser extent the unpleasant memories i n general, were rated as: (1) less clear and v i v i d ; (2) less visually detailed; (3) less likely to occur i n a meaningful order; (4) less well-remembered; (5) less talked about; (6) less frequendy recalled, voluntarily or involuntarily; (7) fewer sensory components including smell, touch, and taste; and containing slighdy less re-experiencing o f the physical sensations, emotions, and thoughts than were present i n the original incident (p. 430). Further, rape memories were less clear and v i v i d , less likely to occur i n a meaningful order, less well-remembered, and less ruminated and spoken about than unpleasant events i n general. K o s s et al.'s (1996) investigation provides evidence that subjective memories for traumatic and unpleasant events i n general are qualitatively different f rom other types o f autobiographical events, according to the people recalling them. H o w e v e r , the qualities o f these memories were not available for assessment by external judges (participants were asked to gauge their memories for the existence o f the criteria). That is, i t is not k n o w n i f free narrative reports about the negative experiences might have contradicted the subjective The Architecture of Childhood Memories 100 perceptions for the memories. G i v e n the generally poor memories reported, it may be that some o f the participants had experienced dissociative symptoms at the time o f the events (e.g., F o a & Rothbaum, 1998). In any case, this study yielded findings requiring further exploration. In this dissertation, the unpleasant memories themselves (including false ones) were provided to the experimenter for both an objective analysis and the subjective characteristics o f the memories. In summary, questions continue regarding the nature o f the adults' memories for stressful events. However , taken alone, field research (e.g., Christ ianson & Hubinette , 1993; Wagenaar & Groenewed, 1990; Yu i l l e & Cutshall , 1986) consistendy reveals that negatively-valenced events can be reconstructed i n an accurate, detailed, and temporally-consonant fashion. The associated memories can also be highly resistant to the distorting effects o f misleading verbal details. Central information is generally reconstructed wel l , as stress tends to focus attention o n these aspects at the time o f encoding. However , this research is l imited to events w h i c h occurred during adulthood. It is not clear whether reconstruction ability w o u l d show a similar pattern f rom chi ldhood into adulthood, or the degree to w h i c h distortion w o u l d play a part i n the quality o f the memories. O n the other hand, the event's meaning w o u l d likely change over the course o f this period and the "current mental structures" w o u l d undoubtedly evolve w h i c h might affect aspects o f the recollection. The finding o f K o s s et al. (1996) that people report memories for unpleasant (including potentially traumatic) events to be o f " p o o r e r " quality than other memories remains to be explained. Quite recendy, researchers have begun to look at adults' reconstructions o f chi ldhood events. Read (1997) found that about 2 0 % o f respondents to his survey indicated that they had significant gaps i n their memory begmning after age 3. Surprisingly, sixty percent respondents reported that they had forgotten about a series o f related The Architecture of Childhood Memories 101 experiences for an extended period o f time. Read also found that participants often retrospectively reported forgetting and recovering memories for various ordinary and emotional events, inc luding sexual abuse. This suggests that even emotional events can be forgotten i n some people. T h e Reconstruct ion o f E m o t i o n a l Events by Chi ldren T h e investigation o f memories for ch i ldhood emotional experiences has l o n g been a concern i n the psychoanalytic tradition (Erdelyi, 1990; H y m a n et al., 1995). T h e capacity o f children to recall events they have witnessed or experienced has also been the subject o f a great deal o f research over the past twenty years (e.g., see C e c i & Bruck , 1993; D o r i s , 1991; G o o d m a n & Bottoms , 1993). O n the question o f created/implanted memories, one issue is whether reconstructive processes occurring i n ch i ldhood w i l l render emotional memories distinctive f rom memories for false or fabricated chi ldhood events created later i n adulthood. Studies o n memory development (e.g., F ivush & H u d s o n , 1990; N e l s o n , 1993) indicate that children have good encoding and retrieval (i.e., learning) capacities for impl ic i t memory f rom early o n whereas retrieval strategies for explicit memory o f experienced events may be less proficient relative to adults (Siegel, 1995). In the present eyewitness era, numerous studies have been conducted comparing the reports o f chi ld and adult witnesses to a c o m m o n event (e.g., G o o d m a n & Reed, 1986; M a r i n , Ho lmes , G u t h , & K o v a c , 1979; Parker, Haverf ie ld , & Baker-Thomas , 1986). Such studies typically investigate differences primarily o n three criteria -amount, accuracy, and incorporat ion o f misleading information. T h e general pattern o f results is that the memory reports o f young children contain less information than adults but w i t h a similar rate o f accuracy i f appropriate questioning techniques were used (typically exceeding 80% i n free recall testing) (e.g., Y u i l l e , 1988). The Architecture of Childhood Memories 102 However , young children are more susceptible to the effects o f misleading informat ion i n their reconstructions o f events compared to older children and adults (Brainerd & Ornste in , 1991; C e c i & B r u c k , 1993; C e c i & Le ichtman, 1992; G o o d m a n & Reed, 1986; Y u i l l e , 1988), although the degree o f heightened suggestibility is still a matter o f controversy (e.g., see C e c i & Bruck, 1993; G o o d m a n & Clarke-Stewart, 1991). A s summarized by Batterman-Faunce and G o o d m a n (1993), young children have been more influenced by leading questions w h e n asked for descriptions o f unfamiliar people rather than events (e.g., G o o d m a n et a l , 1991), w h e n pressed for additional information (e.g., Dent , 1990), after a lengthy delay, w h e n coerced, w h e n questioned by an authority figure, and wi th repeated questioning. A s wel l , children's suggestibility appears to be the greatest w h e n memory for the original event is poor and the mis information pertains to peripheral details (Yuille, 1988). Some commentators have criticized the over-concentration on children's suggestibility i n the literature (Marxsen, Y u i l l e , & Nisbet , 1995), reiterating Wel l s ' (1978) recommendation that forensic researchers focus o n remedial actions to min imize the problems associated w i t h suggestion early on . Nonetheless , it is clear that young children are relatively suggestible, i f the above factors are present (Ceci & Bruck , 1993; Cec i , Crossman, Gilstrap, & Scull in, 1998). Presumably, distortions i n emotional event reconstructions w h i c h occurred during ch i ldhood could remain i n the memories into adulthood. O n the other hand, under certain circumstances, the distortion might be eradicated i f the accurate vers ion o f the event were repeatedly communicated to the person (or for that matter a different inaccurate version) or i f the distorted reconstruction no longer matched the person's schemata, semantic knowledge, etc. In the present study, adults' recall for stressful ch i ldhood events (likely about a decade or two before the interviews) was being investigated. Should emotional memories created during chi ldhood be different f rom those created i n adulthood? There are a few factors w h i c h may The Architecture of Childhood Memories 103 contribute to qualitative differences between memories for events processed i n ch i ldhood and adulthood. There is some agreement among developmental psychologists that basic processes o f remembering evolve f rom chi ldhood to adulthood. Y o u n g children appear to be less proficient at processing information because o f attentional factors, l imited memory capacity, and l imited memory strategies (Siegler, 1986; Y u i l l e , 1988), and their abilities to reconstruct life events may be different. At tent ional capacity Chi ldren exhibit l imited attentional capacity i n that they are less able to attend simultaneously to multiple aspects o f an event. Recognizing this pattern, Plotinus (c. 244 A D / 1962) v iewed it as enhancing children's memories: ' 'Chi ldren are better at remembering; the things presented to them are not constandy withdrawn but remain i n sight; i n their case the attention is still l imited and not scattered; those whose faculty and mental activity are busied u p o n a multitude o f subjects pass quickly over all, l ingering o n n o n e " (p. 108). Fewer details become encoded i n the first place and less information is subsequendy recalled (Joffe, 1992). T h e absence o f certain information (to contradict suggested information) may also enhance suggestibility (Ceci & Bruck , 1993; C e c i et al., 1998) and may be conducive to memory distortion generally. Some research has indicated that children attend better to actions wi th in a nove l event than objects wi th in the event (e.g., Ross & Ne l son , 1986; T o d d & Perlmutter, 1983), relative to adults. Genuine ch i ldhood events recalled i n adulthood, then, may contain fewer references to objects present during the experience compared to memories for ch i ldhood created during adulthood. O n the other hand, many later reconstructions o f a true ch i ldhood event may result i n object constructions or creations, as wel l (as part o f the reconstructive process) and real and created memories could thus be similar i n this regard. Ornstein, Merri t t , and Baker-Ward (1995) The Architecture of Childhood Memories 104 also suggested that individual differences i n behavioral styles such as temperament may determine h o w children respond to stressful situations w h i c h i n turn may influence the attention directed at particular features o f the event. Clearly, the complexities o f attention and memory have not yet been unravelled. L e v e l o f cognitive complexity It is n o w clear that the amount o f previous knowledge a person has for an event or other memory stimulus is direcdy related to h o w wel l it is processed (Chi & Cec i , 1987; C e c i et al., 1998; Schneider & Pressley, 1989). A s wel l , the capacity to share an event narratively w i t h others facilitates the quality o f the memory itself (e.g., N e l s o n , 1993). A s Flavel l (1985) stated, "what the head knows has an enormous effect o n what the head learns and remembers. But , o f course, what the head knows changes enormously i n the course o f development, and these changes consequently make for changes i n memory behavior" (p. 213). Generally, the more elaborate a person's cognitive structures for an event, the better the person is capable o f meaningfully interpreting mcorning information. Personal events containing a h igh emotional component may be more difficult for children to process because they have less experience, direct or vicarious, w i t h such events (it should be noted that there are certain emotional events o f w h i c h an adult may have no knowledge and may also process poorly, such as natural disasters [e.g., M c M i l l e n & Smith, 1997]). F o r certain repeated emotional events (such as abuse) children may come to gain a higher level o f cognitive complexity over time i n w h i c h case memories may be better for later incidents than earlier incidents (independent o f the greater time lapse). T h e lower level o f cognitive complexity i n children results i n less informat ion about a nove l event being recalled and level o f susceptibility to memory distortion f rom other sources (e.g., C e c i & Bruck , 1993; K i n g & Yui l l e , 1986). O n the other hand, for some types o f memory stimuli , The Architecture of Childhood Memories 105 children may have a higher level o f cognitive complexity and exhibit superior recall a n d / o r greater resistance to leading questions than adults. F o r example, children exhibit superior memory for cartoons and experienced chi ld chess players recall game patterns better than inexperienced adult chess players (see C h i & Cec i , 1987). A n d o f course, a higher degree o f pre-existing knowledge and schemata can have negative distorting effects o n processing and later remembering events. It plays a considerable role i n the sometimes erroneous memory reconstruction i n adults. Johnson (1988) provided an iUuminating hypothetical case: Take, for example, the event memory " John said he w o u l d help me but he didn't ." Thi s could be an accurate representation o f an actual prior experience, or it could be a memory for an imagined or anticipated event that never happened. F r o m events one also acquires knowledge, attitudes and beliefs, such as J o h n is unreliable" or (if J o h n happens to be a lawyer) "lawyers are unreliable." I f no corresponding event actually happened, the memory that " John said he w o u l d help but didn't" is a failure o f reality moni tor ing at the event level; the belief that J o h n is unreliable w o u l d be a failure o f reality moni tor ing at the level o f beliefs and knowledge (p. 392). M o s t emotional events i n chi ldhood probably involve novel experiences (one can envision unfortunate exceptions to this statement), for w h i c h the chi ld has litde cognitive complexity. Hence , these events may be subject to considerable distortion over the course o f later ch i ldhood and adulthood as they become better understood by the person. F o r example, i f a five-year-old had a fearful experience being bitten by a miniature poodle, (s)he might later reconstruct a large Rottweiler as the perpetrator given h i s / her particular beliefs/ knowledge about dangerous dogs. The Architecture of Childhood Memories 106 L i m i t e d reconstructive resources A m o n g Piaget's most important observations was that children do not experience their environments i n a passive way but rather actively construct their experiences, i n part because schemata are emerging and developing. Their v iew o f the w o r l d is thus less constrained than adults. H o w e v e r , young children appear to have a circumscribed repertoire o f encoding and retrieval (i.e., reconstruction) strategies compared to adults (e.g., Ne l son , 1993; Schneider & Pressley, 1989). Already mentioned was the f inding that children generally attend to fewer details o f an event i n the first place w h i c h may l imit their ability to accurately reconstruct the event. B u t chi ldren also seem to be less proficient at accessing informat ion pertinent to reconstructing the event. This may be less problematic i n contexts similar to the encoding environment. That is, children's memory (especially free recall) seems to be highly context-dependent, benefitting f rom context reinstatement (e.g., Davies & T h o m p s o n , 1988; Pipe , Gee , & W i l s o n , 1993; Porter, 1992; W i l k i n s o n , 1988), i n accordance w i t h the encoding-specificity principle (Tulving & T h o m p s o n , 1973). Context-reinstatement, a critical aspect o f the Cognit ive Interview (see Geise lman & Fisher, 1989; Fisher & Geiselman, 1992), probably facilitates the amount o f informat ion (with no decrease i n accuracy) recalled by children (McCauley & Fisher, 1995) via this interview format. A va l id ch i ldhood memory report generated i n adulthood may correspond more closely to a memory report for the same event generated during chi ldhood than a memory report for an event occurring i n adulthood (a created memory). In summary, infrequent emotional ch i ldhood events may be accurate but abbreviated during chi ldhood but could undergo substantial distortion as the individual develops greater cognitive complexity for the event category. I f litde distortion has occurred, the adult may recall m u c h o f the event f rom a child's developmental perspective (e.g., I thought the dog didn't like what I said and so he bit me") . The Architecture of Childhood Memories 107 Context and scripts Adul t s ' and children's memories for personal events may differ i n the level o f contextual informat ion present, as recognized by the Statement Analysis approach to credibility assessment (e.g., Porter & Y u i l l e , 1996). A c c o r d i n g to Cole and Cole (1988), the weakest aspect o f young children's' memory is "locating events i n time and being able to give specific examples o f recurrent events" (p. 321) w h i c h may make i t difficult for adults to piece together their accounts into a coherent narrative. Thus , fabricated memories or created memories emerging i n adulthood may contain a higher degree o f contextual information than memories for events encoded i n early chi ldhood. O n the other hand, reconstruction and modif icat ion o f the event over the years into adulthood may render the ch i ldhood memory similar to the other types (Belli & Loftus , 1995; Lindsay & Read, 1994; Loftus , 1993, 1997a). Further research is required to resolve this issue. A n o t h e r f inding is that children tend to recall scripts better than specific autobiographical events (e.g., C e c i et al., 1998; F ivush, 1984). Script memory is only a relevant consideration here i f the emotional event being recalled was recurrent and not novel . A l t h o u g h discussions o f script memory tend to focus o n c o m m o n events such as dining i n restaurants, scripts can also develop for repeated criminal events such as ch i ldhood abuse (Yuil le , 1988). However , children tend to recall a novel episode wel l , especially emotional ones, i n terms o f temporal sequence (Ceci et al., 1998; F ivush , H u d s o n , & Ne l son , 1984) and the central details o f the event (Myers & B l u h m , 1985; Smith, Ratner, & Hobart , 1987). It w o u l d appear that script memory can distort children's recall o f their experiences (Kai l , 1990). I f some events occur i n a different order f rom the sequence specified by an established script, the events may be forgotten or recalled i n the scripted order (Hudson & Ne l son , 1983). A s wel l , chi ldren may The Architecture of Childhood Memories 108 sometimes have difficulty i n distinguishing what they actually experienced f rom what was specified by a script (Hudson, 1988). O f course, for a nove l emotional event, the corresponding scripts w i l l be lacking (Nelson, 1989). Recal l for these events w i l l not be subject to the same distorting effects as scripted events. I f the emotional event recurs, children may recall the init ial event i n great detail but the subsequent experiences w i l l be recalled i n terms o f generalities (Hudson & N e l s o n , 1986). In an investigation o f the memories o f twenty children w h o had suffered various emotional experiences before the age o f five, Terr (1988) found that repeated painful events are less fully remembered than are single episodes although accuracy appeared to be unaffected by frequency o f the event. In terms o f the nature o f the events, Terr found that quality o f the memories was not related to the type o f event (sexual, accidental, witnessed, etc.), casting doubt o n the view that sexual abuse is more poorly recalled, or more often repressed, than other emotional events. In terms o f created memories, it may be the case that a person w h o has experienced a series o f events similar to the one being suggested may be more susceptible to memory implantation (e.g., Marxsen et al., 1996). Language and the advent o f photography Ta lk ing about memories for events can have a strong effect o n the quality o f memories i n bo th ch i ldhood and adulthood. A s pointed out by Siegel (1995), "memory talk," i n w h i c h parents converse w i t h their children about their memories, influences the development o f autobiographical memory. Such social interaction can be reinforcing for children (and adults) and teach children h o w to "remember" w i t h narratives. Ta lk ing about events can improve reconstructions for them, especially the core details, but can also progressively distort the memories w i t h each retelling, as Bardett (1932) had suggested. O f course, telling a personal The Architecture of Childhood Memories 109 story can be engaging for the recipient and the rememberer can pick up subde clues to w h i c h aspects o f stories are more or less likely to be rewarded w i t h interest, sympathy, intrigue, and so on , evidenced i n the case o f the K i l l e r Tomato attack. Photography has probably contributed to the evolution o f autobiographical memory, as well . O n l y i n the last century have most children grown up repeatedly v iewing exact replicas o f meaningful experiences f rom their past. T h e consequences o n memory o f this interesting turn o f events are unclear but it likely enhances the clarity o f the memories for the photographed events, while simultaneously distorting them. That is, i f we repeatedly see ourselves i n a photo and discuss the event i n this observer perspective, we may eventually recall it that way (seeing ourselves i n the scene), w h i c h w o u l d offer pr ima facie evidence for a distorted reconstruction. N o t only does a photograph say a thousand words but it generates a good many reflections. Husser l (1966), one o f the only philosophers to consider the potential effect o f photography o n cognit ion, posited that exact replicas o f life events serve to improve our (reconstructive) memories by impell ing us to "overcome the kinds o f errors and illusions proper to m e m o r y " (Husserl, 1966/1974, p. 98). O f course, most negative events that happen to children are not photographed or commonly discussed so i t is the positive events w h i c h are likely to be most subject to this type o f distortion (more positive events may also become distorted through nostalgia). Naturalistic studies o f children's reconstructions A new and intriguing line o f research on children's memory aimed at increasing ecological vaHdity has been conducted i n w h i c h children's recall for naturalistic stressors is tested. Children's memory for invasive and often traumatic medical procedures has been looked at (e.g., C e c i et al., 1998; G o o d m a n , Hi r schman, Hepps , & Rudy, 1991; Peters, 1988; Peterson, The Architecture of Childhood Memories 110 1996; Peterson & B e l l , 1996). F o r example, Peterson (1996) found that young N e w f o u n d l a n d children were able to recall the central details o f emergency surgical procedures up to six months after the event w i t h "extreme accuracy" (e.g., 94% accuracy rate i n 4-year-olds six months after the procedure). G o o d m a n et al. (1991) examined children's memory for an inoculat ion procedure, f inding that up to a year after the shots, the children were highly resistant to misleading questions w i t h abusive thematic content (e.g., " D i d she hit you?", " D i d she kiss you?"), rarely replying i n the affirmative. A n d , importandy, the children w h o had been most stressed during the procedure recalled the event more completely w h e n interviewed a few days later than those w h o appeared to have been less agitated. Peterson and B e l l (1996) found that for children w h o had suffered stressful injuries (defined as " traumatic" by hospital r o o m staff), distress at the time o f the injury was unrelated to h o w wel l i t was recalled whereas distress dur ing hospital treatment decreased amount recalled. Children's reconstructions for cr iminal vict imizat ion experiences have also been explored. Terr (1979) studied 26 children aged 5 to 14 years w h o had together experienced an extremely stressful event - being kidnapped and held hostage o n a schoolbus. They spent the final sixteen hours o f the ordeal i n a dark "ho le " i n the ground (a buried truck trailer). W h e n the children were interviewed five to thirteen months after the event, all were able to provide v i v i d and accurate reports o f the kidnapping, including their o w n and peers' behaviors, a "fully detailed story o f what happened." Importandy, none o f the children exhibited any sign o f memory impairment or "the repression, amnesia, memory lapses, emotional numbing , or blurr ing o f consciousness described i n adults after extreme stress" (p. 564). It appears that for this type o f v ict imizat ion experience, the children's memories were enhanced (or at least not impaired) by the stress o f the situation over the study period, although a variety o f other psychological effects such as persistent fears and panic attacks were present (Terr, 1979). N o The Architecture of Childhood Memories 111 doubt, there was m u c h communicat ion among the children during the event-study inter im w h i c h may have facilitated the detailed stories. T o summarize, controlled research indicates that young children tend to recall emotional events w i t h a degree o f accuracy comparable to older children and adults. H o w e v e r , they offer less informat ion and display a higher susceptibility to the distorting effects o f mis informat ion i n the reconstruction o f such events. This elevated suggestibility is most pronounced w h e n questions involve unfamiliar people, w h e n pressed for additional information, after lengthy delays, w h e n coerced, w h e n questioned by an authority figure, and w i t h repeated questioning. Deficits i n attentional capacity, cognitive capacity, l imited reconstructive processes, and other factors contribute to variations i n patterns o f event reconstructions f rom adults. Reconstructions o f emotional ch i ldhood events w i l l likely be subject to some distortion f rom ch i ldhood to adulthood because o f changing current mental structures. T h e meaning o f the event w i l l change for the person and multiple inter- and intra- personal reflections may result i n the type o f distortions described by Bardett. Naturalistic studies indicate that children can reconstruct highly stressful events w i t h some degree o f accuracy, completeness, and resistance to suggestion. I f the emotional events f rom chi ldhood being recalled by adults have been reflected o n often over the years, the memories might be expected to be coherent (given the tendency for coherent narratives to evolve w i t h re-tellings), v iv id , relevant (given the well-thought-out meaning attached to the event), and w i t h some memories for the original cognitive experiences (e.g., " I remember wondering why the nice dog bit my arm"). The Architecture of Childhood Memories 112 T h e Reconstruction o f "Traumat ic " Events W h i l e some doctors never trouble their heads about traumatic memories, and are not even aware o f the fact that they exist, and whereas others fancy them everywhere, there is a place for persons w h o take a middle course, and to believe they are able to detect the existence o f traumatic memories i n specific cases (Janet, 1919/ 1925, p. 670). A s the literature review to this point has conveyed, there have been contradictary findings o n the effect o f significant stress o n memory, although ecologically val id studies suggest a facilitatory role. Some have argued that certain emotional events are experienced and remembered i n a special manner - traumatic events, or "traumata" (e.g., Christ ianson, 1992; K o s s , T r o m p , & Tharan, 1995; Krysta l , Southwick, & Charney, 1995; Porter & Marxsen , 1998; van der K o l k , 1996, 1997a). Thus , whether an event is experienced as "stressful" or " traumatic" could have implications for the quality o f memories associated w i t h it. Ear ly i n this paper, it was argued that we tend to spend more time and effort reconstructing events w h i c h have the greatest personal relevance. Trauma, as defined below, may represent a pinnacle o f personal threat and might be expected to evoke m u c h reflection and consideration. A s discussed, the reconstruction o f extremely stressful events can be highly accurate as i n the case o f the Chowchi l l a children, or completely impeded as i n very rare cases o f dissociative amnesia (e.g., F o a & Rothbaum, 1998; Porter et al., submitted for publication). A l t h o u g h this bi-polar quality is not wel l understood (Earleywine & G a n n , 1995), one possibility is that central aspects o f traumatic events are generally well-remembered but certain individuals are predisposed to dissociative processes. Dissociative amnesia for traumatic events is thought to result f r o m a number o f psychological processes, inc luding dissociation (Spiegel & Cardena, 1991) and other individual differences. In the present study, adults were interviewed about emotional ch i ldhood events w h i c h may have been traumatic for some and an exploration o f the qualities o f traumatic memories is necessary. The Architecture of Childhood Memories 113 Histor ica l observations o f traumatic memories O n e o f the first psychologists to expkcitly consider the effects o f traumatic events, Pierre Janet described the psychological processes i n w h i c h organisms react to overwhelming psychological trauma and the effects o f trauma on memory (Janet, 1889). H e contended that trauma "narrows the field o f consciousness" (according to Janet, this was a process similar to Freud's repression concept [Ellenberger, 1970]), similar to the modern findings o n stress and attentional focussing, and wrote o f the dissociation o f traumatic memories f rom "the totality o f the sensations and the ideas w h i c h comprised the subject's personality" (Janet, 1925, p. 674), and that such memories could be recalled i n non-verbal form. V a n der K o l k (1996) has fo l lowed i n the footsteps o f Janet i n both his conception o f dissociation and the re-experiencing o f traumatic events as " b o d y memories ." Before Freud, Janet advocated the importance o f recovering traumatic memories for therapeutic reasons but, unlike Freud (in his early writings), also recognized the basic reconstructive nature o f memories for events. Janet and later theorists believed that such memories can be distorted by associated experiences and emotional state at the time o f recall (Janet, 1889; van der K o l k , 1997b; van der K o l k & Fisler, 1995; van der K o l k & van der Hart , 1991) but that such memories generally remain v i v i d and have temporal stability (van der K o l k , 1996). Janet (1928) adopted the Husserlian view that memories are constructed narratives i n w h i c h informat ion is incorporated into an existing knowledge base. A t the end o f the nineteenth century, Janet's idea o f reconstructive memory was atypical, as experimental psychologists adopted Ebbinghaus ' quantitative approach while the psychoanalysis o f F reud and colleagues - w i t h a storehouse approach to traumatic memories - began to dominate psychiatry. Breuer and Freud (1957) argued that, without an oudet, or catharsis, the affect remains attached to the memory w h i c h remains i n its original state: The Architecture of Childhood Memories 114 A t first sight it seems extraordinary that events experienced so long ago should continue to operate so intensely - that their recollection should not be liable to the wearing away process to w h i c h , after all, we see all our memories succumb. . .The fading o f a memory or the losing o f its affect depends on various factors. T h e most important o f these is whether there has been an energetic reaction to the event that provokes an affect (p. 8). This idea that affective inhibi t ion serves to sustain the quality o f autobiographical memories has not been empirically supported. Rather, f rom a reconstructivist perspective affective reflection or re-experiencing o f the event w o u l d enhance the subjective quality o f the memory (e.g., Christ ianson, 1992). T h e recent prob lem o f over-extending " t rauma" In the current age o f the v ic t im, " t rauma" has become a catch-phrase for just about any negative experience conceivable. People are traumatized by events f rom their chi ldhoods, their relationships, accidents, cr iminal victimizations ranging i n gravity f rom the apparendy petty to the serious, the physical and psychological conditions o f the workplace, another person's off-colour comments , and so on. Thus , trauma is probably overdiagnosed and may be rare for most events. M a n y researchers have not been immune to this pattern o f overextension, apparendy assuming that a particular type o f event w i l l be universally experienced as " traumatic ," rendering interpretation o f m u c h o f the literature o n trauma and memory difficult (Porter & Marxsen, 1998). Terr (1979) tided her paper o f the twenty-six kidnapped Chowchi l l a children a study o f "psychic trauma." Unquestionably, this experience was highly stressful for al l the children involved but at what point does highly stressful become traumatic? A c o m m o n and parsimonious definition o f trauma is an "inescapably stressful event that overwhelms people's existing coping mechanisms" (van der K o l k , 1996, p. 279). Traumatic The Architecture of Childhood Memories 115 events represent a special class o f extreme stressful events mainly i n terms o f severity, (approximately 20-30% o f such events may result i n Post-Traumatic Stress Disorder 1 ) (Foa & Rothbaum, 1998). That is, trauma represents the vertex o f response patterns to stress. Breuer and Freud (1893/ 1957) felt that there was a l o w threshold o f when psychic stress becomes psychic trauma. A n y distressing event eliciting an affective response - "fright, anxiety, shame or physical p a i n " - cou ld effect trauma depending o n the susceptibility o f the person. They further maintained that a series o f partial traumas can have a traumatic effect by cumulative effect, or " s u m m a t i o n " (p. 6). N o t e the ^determinate elements o f the definition o f trauma. A t what point does the event "overcome existing coping mechanisms", especially given that psychology has yet to develop a reliable assessment measure o f existing coping mechanisms? A g a i n using the Terr (1979) study, were all o f the children's existing coping mechanisms overcome? G i v e n the inter-personal variation o f h o w people deal w i t h life stressors, probably not. Some individuals are remarkably resilient i n the face o f inescapable and extreme stress. Consider the fo l lowing anecdote about an event that w o u l d universally be considered highly stressful and is, o n the surface, " traumatic ." The Vancouver Sun (October 5, p. A12) documented the abduction o f seven-year-old Kr i s t ina Jacobson f rom a day-care center i n Roseburg, Oregon . The kidnapper, Lance Alexander, took her at gunpoint and led police o n a chase at speeds up to 1 8 0 k m / h r while mutual shooting occurred. Finally the car fl ipped and Alexander negotiated w i t h police for one hour f rom the car before a police sharpshooter ki l led h i m w i t h a single shot. Kr i s t ina was described as "unfazed" by the ordeal remarking, " I can still picture h i m right i n m y head w h e n he died. F r o m here and up the head was off." Af ter detailing her experience to reporters, she added that she had used her B a m b i book to smash a w i n d o w after the man was kil led because she didn't want b l o o d to get o n her sun dress. Undoubtedly , many people -children or adults - w o u l d probably have been "traumatized" by such an extreme stressor. Thi s The Architecture of Childhood Memories 116 is not meant to say w i t h certainty that Kr i s t ina was not traumatized. E m o t i o n a l n u m b i n g and adopting a detached objectivity is coping strategy for trauma i n some individuals (e.g., F o a , Riggs, & Gershuny, 1995). T h e pattern o f cognitive and behavioral responding i n any situation is the result o f a complex person/event interaction, as highlighted by Walter M i s c h e l (1969). Stressful experiences are not unique i n this respect; the experience and memory for it result f rom the interaction o f the individual and the stressful event. Trauma is dependent o n the cop ing resources a particular person brings to a particular stressful situation. T h e person (usually the v ic t im 2 ) , must cognitively appraise the situation as being inescapable and stressful before h i s /her psychological coping mechanisms can be either effective or " o v e r c o m e " (Porter & Marxsen, 1998). In other words , the person must be aware o f the negative and uncontrollable aspects o f an event. A s a consequence o f these conditions, few, i f any, events w i l l invariably be experienced as traumatic but many w i l l usually be stressful, despite a c o m m o n assumption i n discussions o f memories for false events (Rubin, 1996); some negative experiences w i l l me more conducive to the induct ion o f trauma than others, especially rape (e.g., F o a & Rothbaum, 1998). Pervading m u c h o f the clinical literature is the idea that sexual abuse is inevitably experienced as a traumatic event, a v iew exemplified i n a misleading statement by Blake-White and K l i n e (1985): "mos t vict ims o f sex abuse have very few memories o f the actual incest," (because o f dissociation, repression, etc.) (p. 396). A s the subject o f testimony, some events involve an extremely high degree o f personal significance, such as a rape, frequendy experienced as traumatic (e.g., F o a & Rothbaum, 1998), while others h o l d very l o w personal significance, such as witnessing a m i n o r automobile accident. M o s t testimony given i n court w i l l probably relate events falling somewhere between these two poles, such as recalling a v ict imizat ion o f a n o n -violent robbery, witnessing a physical assault, or experiencing a more " m i n o r " sexual assault The Architecture of Childhood Memories 117 (e.g., fondling). T h e personal significance o f an event is contingent o n many factors, all o f them encompassed by h o w a person appraises the situation during and after the event. F o r example, a sexual assault w o u l d not be experienced as traumatic unless the v i c t im appraised it as w r o n g a n d / o r painful at the time it happened (one can envision a case i n w h i c h a young chi ld sees sexual contact w i t h a parent as " n o r m a l " behavior). This appraisal process may be an important consideration i n deterrnining the quality and endurance o f a memory. M o s t witnesses and victims testifying i n court probably have not been traumatized i n this sense. W i t h i n the Canadian justice system, "sexual assault" encompasses a wide range offenses f r o m the relatively minor to extremely serious offenses. Whereas a significant proport ion o f rapes may be experienced as traumatic, a chi ld w h o has been sexualized f rom an early age i n the context o f a non-violent incestuous relationship w o u l d probably not experience trauma. I f the molestation continued into later chi ldhood, the same sexual contact might be re-appraised i n l ight o f the child's progressive knowledge about the wrongfulness and violat ion inherent i n such a relationship and trauma might be experienced (Porter & Marxsen, 1998). That is the person's current mental structures w i l l change the nature o f the memory. G u i l t may be a factor i n the etiology o f rare psychogenic amnesia unrelated to state-dependent memory or dissociation at the time o f the traumatic event (e.g., K i h l s t r o m et al., 1993). T h e literature suggests that events to w h i c h n o feelings o f responsibility by the v i c t i m are attached (e.g., accidents, concentration camps, other shared ordeals) are not l ikely candidates for this type o f memory impairment. G u i l t is a c o m m o n affective experience for vict ims o f certain types o f crimes such as rape, spouse abuse, and ch i ldhood sexual abuse (e.g., F o a & Rothbaum, 1998; K u b a n y , A b u e g , Owens , Brennan, K a p l a n , & Watson , 1995), as w e l l as combat soldiers and certain cr iminal offenders, such as murderers. The main point here is that researchers interested i n the qualities o f reconstructed traumatic events must first establish the existence o f The Architecture of Childhood Memories 118 trauma rather than simply assuming its validity based o n the pre-existence o f a certain type o f event. Particular types o f ch i ldhood events may be appraised as traumatic or stressful i n different individuals. F o r example, the events explored i n the present study invo lved personal injuries, animal attacks, getting lost, and medical procedures, w h i c h w i l l have affected individual participants very differently but probably did not have an aspect o f guilt. W r i t i n g o n the relationship between emotion and memory, Augustine observed that the emotion present while experiencing an event " c lung to his m i n d " so that he could easily remember it. E m o t i o n a l arousal appears to influence perception (e.g., L e D o u x , 1992), attentional focus (Christianson, 1992; Yu i l l e & Daylen , 1998; Yu i l l e & Tollestrup, 1992), and the frequency o f discussion about the event w i t h others (Bohannon & Symons, 1992). A s earlier mentioned, stressful events tend to focus attention (Easterbrook, 1959), so that central details w i l l be recalled we l l and peripheral details poorly (e.g., Christianson, 1992). Pol ice officers w h o have ki l led i n the line o f duty (which may or may not have resulted i n trauma) appear to recall the event v iv id ly but idiosyncratically, i n that the event seems to occur i n s low m o t i o n , through tunnel v i s ion , and to the exclusion o f peripheral information. Parent (1996) interviewed thirty-four such officers i n Bri t i sh C o l u m b i a and found that reconstructions were influenced by temporal, visual, and auditory distortions: " A s the incident unfolded, individual officers noted that their deadly force encounter appeared to occur i n slow mot ion . Of ten their v i s ion was focussed u p o n the perceived threat w i t h rrunimal awareness o f the events taking place around them. Finally, w h e n shots were fired, they were generally heard as muff led sounds, even though the officers were not wearing ear protection devices" (p. 101). The "weapon-focus" effect - that often a weapon is often focussed on during a vict imizat ion to the detriment o f memory for other aspects o f the event - probably has a similar origin. The Architecture of Childhood Memories 119 Memories for most stressful events appear to be consistent over long time periods and subjectively v i v i d , unlike memories for other autobiographical events. Traumatic events show an interesting alternation between re-experiencing and avoiding trauma-related memories (e.g., B r e w i n , Dagleish, & Joseph, 1996; F o a & Rothbaum, 1998). The trauma-related memories appear rapidly and spontaneously, consisting o f v i v i d imagery accompanied by high physiological arousal ("flashbacks"). It has been argued that trauma-related memories for events that happened i n the distant past may be more fragmentary and consist o f isolated visual, auditory, olfactory, or tactile experiences (van der K o l k , 1997a, 1997b; van der ko lk & Fisler, 1995), although this remains a highly contentious issue. M e m o r y impairment sternming f rom psychogenic (as opposed to physical) trauma appears to be exceedingly rare. M o s t sexual assault victims, accident vict ims, witnesses to violence, and so o n continuously recall the event. It is difficult to conceive a more horrif ic or possibly traumatic event than a chi ld witnessing the murder o f a parent and, i n theory, it might represent an " i d e a l " event for psychogenic amnesia to occur. Ye t , Malmquis t (1986) studied sixteen children between the ages o f five and ten w h o had gone through such an ordeal and found that every chi ld held v i v i d recollections for it, even though half o f endorsed the i tem " I wished to banish it f rom my store o f memories , " an i tem o n the Impact o f E v e n t Scale by H o r o w i t z , Wi lner , Kaltreider, and Alvarez (1980). A s i n P T S D , recall for the homic ida l scene recurred unpredictably as flashbacks. T h e memories usually pertained to the wounds infl icted on the parent, but no other information was provided by the author o f specific characteristics o f the memories. There are several naturalistic studies examining the reconstructions o f potentially traumatic events, o f w h i c h some have been discussed. L e o p o l d and D i l l o n (1963) studied 34 m e n w h o had survived an explosion at sea after two ships collided. W h e n interviewed four The Architecture of Childhood Memories 120 years after the accident, the men exhibited symptoms associated w i t h P T S D but none experienced any memory impairment, prompt ing the authors to question the validity o f repression. A l l 25 o f the Chowchi l l a children were able to provide a detailed account o f the kidnapping four years later (Terr, 1979, 1983). Perhaps the most salient example o f individuals w h o experienced a shared horrific event were the concentration camp survivors o f W o r l d W a r II. A n assumption can be made that such an event w o u l d be experienced as traumatic for many victims. Y e t , there is no evidence for widespread memory impairment i n the vict ims; instead, the memories appear to be enduring and w i t h detailed reconstructions, according to Wagenaar and Groenweg's (1990) findings. Recal l that forty years after war, the victims recalled their experiences i n a fashion consistent w i t h their testimony at Nuremberg . Six m e n w h o had testified about horrif ic vict imizat ion experiences at Nuremberg failed to initially ment ion them i n the study. W h e n reminded o f their earlier statements, five o f them soon came to reconstruct the events. Nonetheless, the possibility exists that survivors w h o evidenced memory impairment i n the post-war investigations were not called to testify at Nuremberg i n the first place (and thus w o u l d be exempt f rom Wagenaar and Groenweg's study) 3. Clearly, though, many people vividly recall their traumatic experiences. The quality o f the memories may depend o n h o w the people appraised and dealt w i t h the horrors at the time and whether they were shared. (Interestingly, to m y knowledge, there is no recorded case o f potent memory impairment i n the context o f a shared vict imizat ion experience [i.e., simultaneous multiple victims]). Existential psychologist, V i k t o r F r a n k l developed his system o f logotherapy (Gr. logos: word or thought) based o n his memories for experiences i n a N a z i concentration camp (Frankl, 1949,1969; Johnson, 1966). H e described atrocities i n w h i c h prisoners were able to transcend the torture, pain, and approaching death confronting them. Those w h o were able to transcend the experience often The Architecture of Childhood Memories 121 found a deeper meaning i n life, a capacity for seeking reasons for their existence; the memories became a guiding principle. W h a t about memories for personally-experienced life-threatening accidents or medical procedures? G i v e n their direct relationship to well-being, such events should be reflected u p o n often and reconstruction could result i n a v iv id , generally accurate (with the normative distorting factors considered earlier) memory. This is precisely the case. In one study, 107 people w h o had been treated for severe injuries at an emergency ward sixteen to 51 months before were questioned (Malt, 1988); the only memory impairment evidenced was due to neurological injuries. Similarly, Peterson and B e l l (1996) assessed the memories o f ninety N e w f o u n d l a n d children w h o had been treated for severe or " traumatic" injuries six months earlier. M o s t children recalled the event w i t h a high degree o f accuracy and none experienced any salient memory impairment. Nonetheless, some have argued that trauma may affect the nature o f people's reconstructive processes. V a n der K o l k and van der H a r t (in van der K o l k , 1997a) investigated the relative sensory components o f traumatic memories and ordinary memories i n trauma clinic patients and found that traumatic memories are characterized by a higher sensory component , i n all sensory modalities, prompt ing the first author to conclude that, "people, then, construct a narrative story for a traumatic event out o f the sensory re-enactments." Thi s v iew was foreshadowed by L a n g (1979) w h o argued that emotionally-laden mental images are accompanied by increased autonomic activity. T h e emotional memories are stored i n "associative networks , " consisting o f the sensory elements o f the emotional experience and re-activated w h e n a person is confronted w i t h situations w h i c h stimulate a sufficient number o f sensory elements o f the network. F r o m this perspective, sensory reinstatements precede the verbal reconstruction o f an event instead o f the other way around. Similarly, according to The Architecture of Childhood Memories 122 Johnson and colleagues' (e.g., Johnson et al., 1988) reality monitor ing research, memories for real events contain more sensory informat ion than the imagination. A l t h o u g h it was earlier stated that stressful events cannot be generated i n the laboratory, they have been re-instated i n this context for systematic study i n trauma victims. Recendy, neuro-imaging researchers have evoked traumatic memories o f P T S D sufferers. Neuropsychologica l evidence suggests that emotional events are processed differendy f rom neutral events. There are n o w some early indications o f h o w emotional events may be processed i n the brain. O n e brain component w h i c h appears to be invo lved i n traumatic recollection is the l imbic system, understood to be a primitive system primarily engaged during the four great existential " F ' s " - fleeing, fighting, feeding, and reproduction (van der K o l k , 1997a). Hemispher ic functional differences relevant to autobiographical memories have also been detected. T h e " ra t iona l " left hemisphere is considered the "generative" half o f the brain, invo lved i n the perception and generation o f symbolic representation (van der K o l k , 1997a, 1997b). A stimulus is broken d o w n into categorical elements and constructed into a nove l image i n l ight o f existing schemata. In contrast, the right hemisphere appears to be invo lved i n the expression o f and interpretation o f non-verbal communicat ion (facial expressions, tone o f voice, etc.), the attachment o f emotional information to stimuli, and other basic psychological processes (van der K o l k , 1997b). The neuro-imaging studies have found litde left hemispheric activity, inc luding Broca's area, most associated w i t h transforming experience into speech. Thi s f inding suggests that the events may not be "reconstructible" verbally because they are dissociated f r o m language. A t the same time, areas i n the right hemisphere, believed to process intense emotions and visual images, showed significandy increased activity. T h e amygdala appears to be invo lved i n the control o f a variety o f emotional functions (e.g., L e D o u x , 1992). A c c o r d i n g to some researchers (e.g., van der K o l k , 1997a, 1997b), traumatic events are stored The Architecture of Childhood Memories 123 initially i n sensory f o r m rather than narrative f o r m and can remain inaccessible verbally because the memories are processed i n the right hemisphere only as sensory or bodi ly "f lashbacks." The events eventually can gain verbal representation through rational reconstruction, occurr ing primarily i n the left hemisphere, especially the frontal cortex. V a n der K o l k (1997a) quoted one patient w h o referred to the non-verbal quality o f her traumatic memory: " A l l these intrusive recollections only have an experiential quality to them -1 get lost, confused, and hear things. W h e n I have these flashbacks they are not explainable." However , there are a host o f problems i n the interpretation o f these findings although few have yet been voiced. A n assumption o f van der K o l k and colleagues appears to be that the traumatic memory had never been previously recalled i n verbal form. It is not implausible to suggest that the patients i n the P T S D studies had at one point , maybe soon after the experience, recalled it, or knew it had happened and suppressed the memory into sensory form. There is certainly a storehouse quality to the " b o d y memories " idea (i.e., the sensory components o f particular events are stored i n the right hemisphere), but it is an idea w h i c h warrants further study. A l s o , recall the study discussed earlier by K o s s et al. (1996), w h i c h found that the memories o f rape victims for the rape itself contained fewer sensory components (e.g., smell, physical touch, and taste) and less re-experiencing o f the physical sensations compared to memories less stressful experiences. Clearly, these issues need to be explored more fully. I n summary, an exploration o f the qualities o f reconstructed traumatic events was relevant i n the present context because any negatively-valenced ch i ldhood event can be experienced as " traumatic" by particular individuals. Memories for val id stressful and traumatic events appear to be similar i n certain respects; they remain subjectively v i v i d and can often be accurate. T h e flashbacks resulting f rom traumatic events are unique i n visual clarity and The Architecture of Childhood Memories 124 affective re-experiencing. A l t h o u g h often accurate, like memories for other stressful events, memories for traumata are probably not immune f rom distorting effects o f mis information, despite their subjective durability. Sometimes, they appear to be susceptible to the same distorting forces as other autobiographical memories (Wagenaar & Groenweg , 1991). The Architecture of Childhood Memories 125 CHAPTER FIVE The Architecture of Created Memories The Architecture of Childhood Memories 126 T h e Recent Discovery o f Created/ Implanted Memories A l t h o u g h the impel l ing force for Loftus ' init ial post-event mis information studies was theoretical, widespread application o f the findings i n forensic contexts soon took place (Yuille, 1989; Y u i l l e et al. , 1995; Y u i l l e & Wel l s , 1991). Some psychologists, i n the tradition o f Munsterberg, were wi l l ing to act as experts o n the malleability o f memory, testifying i n al l sorts o f cases i n v o l v i n g witnesses or victims. A n d , taking a quantum leap, some were wi l l ing to testify o n the plausibility o f distorted memory and even false memories for cr iminal events (e.g., sexual abuse) based o n the stop sign result and other post-event mis information research. Inevitably, a heated debate erupted i n the field. Critics condemned such practices as inadvisable o n a number o f points, but, i n particular, the questionable ecological validity o f the research base (see L a w and H u m a n Behavior . 10. 1986. for the published debate), paralleling the scathing critique o f Munsterberg's (1908) O n the Witness Stand i n the Aussage era by A m e r i c a n lawyer, J o h n W i g m o r e (1909). T h e upshot o f this dialogue was the recognition wi th in the field o f the necessity o f conduct ing further research o n memory reconstruction and suggestibility before entering the cour t room (e.g., O l i o , 1994; Leippe, 1995; Pezdek & Roe , 1997; Porter & Marxsen , 1998; Y u i l l e , 1993). I f forensic application was the agenda, the attainment o f ecological validity had to be o n the itinerary. W h e n the false m e m o r y / recovered memory debate began to rage i n the late 1980s, empirical attention m o v e d to extreme memory distortion, whereas the classic misinformation studies had really looked only at the effects o f minor misinformation. There were existing anecdotes but no empirical studies examining false memories for whole events. A n important line o f research to emerge i n the early-1990s, inaugurated by El izabeth Loftus and Stephen C e c i addressed the question o f whether memories for entire false events could be implanted i n subjects i n the laboratory (see Table 3). It is n o w The Architecture of Childhood Memories 127 clear that memories for entire false events can be induced i n a proport ion o f laboratory subjects (see Loftus et al., 1995). Created memories have been implanted i n both children and adults. Created memories i n chi ldhood False memories are by no means rare occurrences i n most o f us. (Wil l iam James, 1890). C r i m i n a l justice systems have l o n g mistrusted the memories o f children, reflecting an enduring belief that children have difficulty differentiating reality f rom fantasy (e.g., C e c i & Bruck , 1993; G o o d m a n , 1984; Leippe & Romanczyk , 1989; Y u i l l e , 1988; cf. Luus , Wel l s , & Turde , 1995). In a recent survey o f judges i n Austral ia , Cashmore and Bussey (1996) found that although the honesty o f chi ld witnesses is rarely questioned, they are still perceived as highly suggestible and prone to fantasy, despite m u c h research showing that they can be reliable witnesses (e.g., see C e c i & Bruck , 1993; Y u i l l e , 1988). A s earlier mentioned, a consistent f inding i n the eyewitness area is that young children are more susceptible to misleading informat ion than older children and adults (e.g., C e c i & Bruck , 1993; C e c i et al., 1998; K i n g & Y u i l l e , 1987) and certain sub-populations o f children may be more susceptible to suggestive pressures than others (such as cognitively challenged children; Dent , 1986, Gudjonsson & G u n n , 1982; and deaf children; Porter, Y u i l l e , & Bent, 1995). However , most o f this research has examined the influence o f leading questions on recall for minor aspects o f witnessed laboratory events (e.g., G o o d m a n & Reed, 1986). The line o f research o f C e c i and colleagues addressed the susceptibility o f children to the creation o f memories for entire false events (Ceci et a l , 1998; Cec i , Crotteau-Huffman, Smith, & Loftus , 1994; Cec i , Loftus , Le ichtman, & Bruck , 1993). The studies showed that created memories could be implanted i n some preschoolers through the provis ion o f misleading information and authoritative social influence. C e c i et al. (1994), i n conjunction w i t h parents, asked 96 children aged three to six years about true and false events The Architecture of Childhood Memories 128 (e.g., getting a hand caught i n a mousetrap and requiring a trip to the hospital) once a week for twelve weeks. They found that about one-third (36% and 32% o f the younger and older children respectively) eventually "recal led" the false events as real, often gready embelhshing the events. C e c i (1995) later commented: W e had n o expectation that this w o u l d result i n the sort o f highly detailed, internally coherent narratives that the children produced.. . .The children w o u l d provide an internally coherent account o f the context i n w h i c h the finger got caught i n the mousetrap as we l l as the affect associated w i t h the event (p. 102). W h e n psychologists w h o specialized i n interviewing children viewed the videotaped memory reports, they were unable to identify w h i c h descriptions o f events were created memories. C e c i concluded that, "repeatedly drinking about a fictitious event can lead some preschool children to produce v i v i d , detailed reports that professionals are unable to discern f rom their reports o f actual events" (p. 103). N e x t , Le ichtman and C e c i (1995) conducted the wel l -known "Sam Stone" study i n w h i c h a stranger named Sam Stone paid a two-minute visit to a daycare center. T h e preschoolers were asked about the visit o n four occasions over the next ten weeks. Some children had heard a stereotype about Sam before his visit (e.g., that he was clumsy) and another group was provided misleading suggestions over a twelve-week period. Forty-two percent o f the children w h o had heard the stereotype and 52% o f the children w h o had received repeated suggestions reported aspects o f Sam's visit w h i c h had not occurred. A g a i n , videotapes o f the children during the final interview were shown to more than a thousand researchers and clinicians w h o w o r k w i t h children; they again failed to detect w h i c h children were accurate or w h i c h events had occurred even though they expressed confidence i n their decision-making (Ceci, 1995). C e c i has argued that there are no features distinguishing real memories f rom The Architecture of Childhood Memories 129 memories for false events, but this may represent a premature observation. Because professionals were unable to identify the fictional aspects o f the memory reports, i t does not necessarily mean that there are n o distinctive characteristics. A l t h o u g h these studies demonstrated the phenomenon o f created memories, some critics held that the creation o f memories for emotional experiences, comparable to sexual abuse, had not been established (e.g., Rudy & G o o d m a n , 1991). In response, subsequent research by C e c i and colleagues (see Bruck, Cec i , Francoeur, & Barr, 1995; C e c i et al., 1998) examined created memories for events such as a painful inoculation, and obtained similar results. However , Pezdek and colleagues (e.g., Pezdek, 1995; Pezdek, Finger, & Hodge , 1996) remained unconvinced that these types o f created memories i n the laboratory are comparable to a chi ld sexual assault experience. The i r experimental attempts at implanting memories i n children for an event more similar to sexual abuse - a rectal enema - had a 0% rate o f success. I n fact, one o f C e c i et al.'s (1994) findings concorded w i t h Pezdek's v iew that more negative memories should be more difficult to implant. C e c i et al. (1994) found that negative events (abusive or direatening experiences) were more resistant to false suggestions than neutral ones (see also, C e c i et al., 1998; Y u i l l e & Cutshall , 1996). Nonetheless, experimental and anecdotal evidence indicates that children have can reconstruct emotional false events i n w h i c h they had not participated (see Loftus , 1993). Haugaard, Repucci , Laurd , and N a u f u l (1991, cited i n Loftus, 1993) led four to seven-year-olds to believe that they had witnessed a man strike a gir l after hearing her lie about the assault. Forty-one children reconstructed the event w i t h contextual information; 39 said it happened near a p o n d , one i n the girl's house, whereas one was unable to specify the location. Pynoos and Nader (1989) investigated children's memories for a sniper attack at an elementary school playground. Some o f the children w h o were interviewed had not been at school at the time o f The Architecture of Childhood Memories 130 the attack, yet still remembered it ; they likely reconstructed the event fo l lowing interactions w i t h children w h o had experienced it. F o r example: O n e gir l said that she was at the school gate nearest the sniper w h e n the shooting began. In truth she was not only out o f the line o f fire, she was hal f a b lock away. A boy w h o had been away o n vacation said he had been on his way to the school , had seen someone ly ing o n the ground, had heard the shots, and then turned back. I n reality, a police barricade prevented anyone f rom approaching the b lock around the school (p. 532). I n summary, i t has been demonstrated that children's reports can be considerably altered i n the face o f preconceived or retroactive misleading information, to the extent that memories for entirely false events have been created i n the laboratory. The extent to w h i c h memories for emotional events can be implanted remains a contentious issue. N o t e that the above anecdote (Pynoos & Nader , 1989) does not involve false reconstructions for a direct, personal vict imizat ion event dike sexual abuse), although some children may have appraised the event as having great personal significance. Created memories i n adulthood It isn't so astonishing, the number o f things that I can remember, as the number o f things I can remember that aren't so - M a r k Twa in . T h e created memory studies w i t h children showed that entire events could be implanted. In light o f the false/ recovered memory debate, the question o f whether adults could come to create memories was also an important one to explore. O n e possibility is that an adult could come to believe (s)he experienced a ch i ldhood event w h i c h , i n reality, d id not occur. Thi s is, o f course, a central issue i n the false m e m o r y / recovered memory debate. Its importance has been highlighted i n an ongoing investigation into alleged widespread ch i ldhood abuse, w i t h 1500 The Architecture of Childhood Memories 131 now-adult complainants w h o had resided i n N o v a Scotia reform schools. T h e u n i o n representing the employees at the institutions claims that up to 90% o f the allegations are "false memories " or fabrications. The other possibility is that adults can creat memories for adulthood experiences. B o t h anecdotes and research o n the susceptibility o f adults to implanted memories has indicated that, as w i t h children, a proport ion o f adults can be convinced that they experienced false events w h i c h never occurred. Nearly a century ago, it was recognized that the memories o f adults for ch i ldhood experiences can be subject to m u c h distortion. In one o f the first empirical investigations o f its k i n d , P o t w i n (1901) asked 100 Yale and M o u n t H o l y o k e College students about their earliest ch i ldhood memories. In interpreting her results, the author noted the ground truth prob lem and was cognizant o f the existence o f memories for events having no objective reality: "Is the memory given truly a memory, or has the person relating it heard the event described so often as to come to believe he remembers it?" (p. 597) 4. In the same year, Seashore (1901) commented, " . . i f you desire faithful testimony f rom a chi ld you must not ask h i m any questions, n o r al low h i m to make an oral report, but require h i m to write spontaneously what he knows. Th i s is especially true about children because they are less capable than adults to distinguish between fact and fiction. B u t h o w many adults are not grown up children i n this respect?" (p. 615). Piaget described a now-classic anecdote f rom his life w h i c h suggested that created memories for emotional ch i ldhood events could occur i n adults. Piaget held a v i v i d ch i ldhood memory o f an attempted kidnapping for more than a decade but later discovered that the event had not actually occurred. H i s nanny admitted that she had concocted the kidnapping story and was experiencing guilt over the watch she had received as a reward for supposedly saving h i m (Piaget, 1962, i n Loftus & K e t c h a m , 1991). In considering this incident, Piaget later stated, The Architecture of Childhood Memories 132 "I...must have heard, as a chi ld , the account o f this story, w h i c h my parents believed, and projected i t into the past i n the f o r m o f a visual memory" (p. 531). Exper imenta l studies o n implanted chi ldhood memories H u m a n memory is a marvelous but fallacious instrument.. .The memories w h i c h lie wi th in us are not carved i n stone, not only do they tend to become erased as the years go by, but often they change, or even increase by incorporating extraneous features -Italian author P r i m o L e v i (1988). Seashore's suspicion that, like children, adults could sometimes have difficulty k n o w i n g whether an experience was real or imagined has been borne out empirically. In the last few years, the first studies on created memories have been conducted (see Table 3). It is n o w clear that memories for false chi ldhood events can be experimentally implanted i n some propor t ion o f adults. H o w e v e r , as Table 3 indicates, both the operational definitions o f " implanted m e m o r y " and the implantation rates have varied. In fact, i n two important studies (Loftus & Pickre l l , 1995; H y m a n et al., 1995) an implanted memory has been defined as occurring when a participant "incorporates false informat ion" pertaining to the suggested event into a memory, a liberal definition. Further, Loftus ' (1997) claim i n Scientific A m e r i c a n that "P lant ing memories is not a particularly difficult thing to d o " (p. 64) has been challenged (e.g., Pezdek & Roe , 1997). Lof tus ' asserti