The purpose of this chapter is to analyze the concept of “flashbulb” memory and contrast it with other forms of human memory. The construct of flashbulb memory was introduced in a seminal paper by Brown and Kulik (1977) to account for memories of events such as the assassination of John F. Kennedy. Brown and Kulik described flashbulb memories as memories for the circumstances of hearing about a highly surprising and consequential event. These memories were said to be like a photograph, to show very little forgetting, and to be produced by a special purpose biological mechanism. Most recent work on this topic derives from the Brown and Kulik paper, and this chapter will begin with a conceptual analysis of that paper. The analysis will examine, in turn, each of the major theoretical and empirical claims of the original Brown and Kulik paper.

Brown and Kulik – theory

Circumstances (news reception context)

The core phenomenon described by Brown and Kulik (1977) is that certain events give rise to memories that show little forgetting. These flashbulb memories include both the central event and the circumstances in which one learned of the event.

Mental imagery

It appears to me that Brown and Kulik believe that the recollection of flashbulb memories involves the occurrence of visual images. Brown and Kulik do not state this explicitly, but it is the only interpretation I can give to their statement that flashbulb memories have a “primary, ‘live’ quality that is almost perceptual” (p. 74).

Commenting on this conference is a hard task. Every clever slant I could think of, and many I could not, were covered by comments in the continuing open discussion. I reflected on what others might do in this situation, and chose Allen Newell as a model. What he did to organize his discussion, on at least one occasion like this, was to spend half of his space on an issue not directly related to the papers of the conference. He had something worth saying that is as appropriate to this conference as it was to his, so I will just use his talk again. After all, he first gave his discussion in 1972 and we have heard here how confused memories can get over time. Newell began as follows.

This is the fourth volume in the Emory Symposia in Cognition series. Like two of its predecessors, Remembering Reconsidered (edited by Ulric Neisser and Eugene Winograd, 1988) and Knowing and Remembering in Young Children (edited by Robyn Fivush and Judith A. Hudson, 1990), it deals with memory. But whereas those earlier volumes were overviews of broad fields, this book is devoted to a single problem and – in large part – a single research paradigm. That problem is the relation between affect and memory; it is studied here by examining recollections of unexpected and emotional events. Although such recollections have only been called “flashbulb memories” since 1977 (the phrase was coined by Roger Brown and James Kulik in a seminal paper), they have long been the subject of theoretical speculation. Recently they have also been the subject of empirical research, much of it focused on memories of one particular event: the explosion of the Space Shuttle Challenger on January 28, 1986. It is those studies that form the core of this book.

The Emory Cognition Project conference on flashbulb memories was held on February 2–3, 1990. Our primary goal was to bring together everyone who had done research on memories of the Challenger episode and to review each others' data and learn from each others' methods. We hoped that the conferees would reach at least a sense of the meeting on the status of such memories: How do they compare with other kinds of recollections?

Several years ago, the teacher of a preschool class in the Boston area held a morning meeting to announce a sad occurrence: The mother of one of the children had died following a long illness. As the teacher began to talk, several children anticipated the topic: “K's mom died”; “I know, K's momdied.” Some children also spontaneously described their own personal circumstances when they originally heard the news: “My mom told me at the hospital”; “My mom told me at breakfast”; “My mom was making orange juice”; “My mom and dad told me”; “My mom said, ‘You won't be able to realize this.’” The teacher explained that K was fine, that K's mother had been sick for a long time, that the event was very unusual, and that it made people feel very sad. The discussion then turned to the general topic of death and the children's fears and misconceptions.

Preschool personnel observed and recorded the emotionally charged discussion, but had they not been aware of the author's interest in memories of personal circumstances, they might well have overlooked the children's comments about how they had “heard the news.” The target event itself – the death of a parent of a young classmate – was terrifying, and needed to be presented and discussed in a supportive context. On the other hand, the children's memories of their own circumstances upon hearing the news evoke the same sense of “mystery” first identified by Brown and Kulik (1977) with respect to adults' “flashbulb” memories of public tragedies.

The two memories above are actual written responses to the question “How did you first hear the news of the Challenger disaster?” The first account was given in the fall of 1988, long after the event, by an Emory senior whom we will call “RT.” It was a vivid recollection, which met or exceeded all the standard tests of a “flashbulb memory.” Asked for 5-point confidence ratings of various aspects of the memory, RT hit the top of the scale: 5 on How did you hear it? (television), Where were you? (her room), What were you doing? (watching television), Who was with you? (roommate), and How did you feel? (shocked and upset).

The chapters in this volume, as well as the discussions at the conference from which the volume emerged, illustrate clearly that the study of “flashbulb” memories raises a wide variety of difficult and often controversial issues. In commenting on the chapters by Larsen (this volume) and Neisser and Harsch (this volume), however, I will focus primarily on a single question: What are the implications of these studies for the hypothesis of a special flashbulb memory mechanism?

The special-mechanism hypothesis

The special-mechanism hypothesis has dominated flashbulb memory research since Brown and Kulik's 1977 article. Larsen addresses this hypothesis directly in his chapter; and although Neisser and Harsch do not couch their discussion in terms of the special-mechanism hypothesis, the data they present are certain to figure prominently in any future discussions of the hypothesis.

The Larsen study and the Neisser and Harsch study report data that will be, and I think should be, taken as evidence against the special-mechanism hypothesis. However, both studies may attract an objection that has been raised against previous critiques of the special-mechanism hypothesis. Specifically, the objection may be that neither the Larsen study nor the Neisser and Harsch study definitively refutes the special-mechanism hypothesis, because neither study demonstrates clearly that ordinary memory mechanisms are sufficient to account for flashbulb memories. This objection deserves careful attention, because it is entirely valid, but at the same time entirely ineffective as a defense of the special-mechanism hypothesis.

The event

On January 28, 1986 at 11:38 a.m. EST, the Space Shuttle Challenger rose into the sky on a pillar of fire. It carried five men and two women aloft, including the first black astronaut, and the first teacher to go into space, Christa McAuliffe. Approximately 72 seconds into the flight, shortly after the astronauts had received and acknowledged the command to go to full throttle, the main external fuel tank exploded. The Challenger disintegrated, killing all on board before the horrified eyes of millions. A high-level executive for Morton-Thiokol, the company that made the solid rocket boosters, watched the vapor trail from the moment of launch and knew something was wrong. After the explosion the booster engines continued to burn and flew off on tangential trajectories. Ironically, his first thought was that at least the boosters that his company manufactured had performed well. He remembered wondering who was going to head the disaster investigation and who was to blame. Shortly thereafter, he was to receive another shock. The booster rockets were responsible for the explosion, and his company's executive decisions were shown to have been determined by expediency and characterized by a blatant disregard for safety.

If the man in question had been only one out of a few people with unusual memories for their personal discoveries of the Challenger disaster, there would be no reason for this book.

Introduction

The goal of this chapter is to describe the biological systems that appear to promote the formation of memories for especially important information. The findings discussed here suggest that the most direct extrapolation from the evidence currently available is that “flashbulb” memories are not unique but represent a special case of more general neuroendocrine regulation of the biological processes responsible for storing information. It is also important to note that the same physiological processes appear to regulate not only memory systems but also a wide range of nonmemory neural and behavioral measures. Thus, at a neurobiological level, the findings suggest a more broadly based behavioral conception of what is regulated–including, but going beyond, memory.

The major point is that information comes in and is acted upon in retrograde fashion by neuroendocrine responses to the experience to promote the storage of recent information. The chapters included in this volume consider the viability of the view that there are unusually complete memories for ancillary information surrounding significant events. These firm memories are usually tested for events important to a large segment of a society, but presumably also apply to individually based significant events as well. It is only the latter case – individually specific memories – that is addressed directly by the neurobiological data.

At a biological level, there are two interacting processes that must be considered to gain a full understanding of the biological bases of memory storage. One consists of the neurobiological substrate mechanisms themselves.

Some more or less ordinary memories

On the evening of March 1, 1986, I was sitting at my portable computer just before going to bed. I was entering the following brief description into a database, adhering to a daily routine I had started on January 1 of that year:

I was in the middle of the peaceful but tedious task of performing an experiment on my own memory. This terse record of a violent and quite upsetting event was part of the material that I would be probed about some months later by the computer and try to remember. Notice particularly that I included information on both the news itself and my personal circumstances, or context, when I first heard the news.

There were other, equally upsetting, events in the database that I recorded during the 6 months of this diary keeping. Just 2 months later, this was entered:

Dates are important in the study of “flashbulb memories.” President John F. Kennedy was assassinated on November 22, 1963 and Brown and Kulik asked people about their memories for that event in 1975 before publishing their famous article in 1977. The 12-year retention interval lent credibility to the claim of remarkable memory; it seemed unusual then, as it does now, for people to recall events with such vividness after so long a time. The flashbulb metaphor has remained controversial since Brown and Kulik coined it, and the explosion of the Challenger shuttle on January 28, 1986 provided an opportunity to examine it more closely. Several enterprising memory researchers independently set out to ask subjects to describe how they had heard the news of the shuttle disaster. Unlike earlier studies about the Kennedy assassination, these studies were undertaken within a short time after the event. Some researchers then waited months or years before questioning the same informants again.

This new wave of research on flashbulb memories has been designed to approach many of the questions raised by Brown and Kulik (1977) and by their critics. Because a number of investigators had been working independently on the same problem, a conference at which they could compare findings and discuss their implications for the relationship between affect and memory seemed appropriate. That conference was held at Emory University on February 2–3, 1990.

“The past is a foreign country … they do things differently there.” So begins L. P. Hartley's classic book, The Go-Between (1953). In thinking about “flashbulb” memories, specifically about the match between their present character and the past experience that gave rise to them, it is tempting to ask a similarly phrased question: Did they do things differently there?

There is no doubt that present memories about past traumatic episodes persist in the minds of most of us. But were things different then (in reality) than they are now (in our minds)? It turns out that some aspects of traumatic experiences do apparently persist quite accurately, whereas other aspects get altered along the way. These persisting flashbulbs, dotted with error, raise a question about the kind of memory system that can accommodate both persistence and error. What functions would such a system serve, asks Pillemer (this volume)? Equally intriguing is why some traumatic events get transformed into persisting flashbulbs, whereas other traumatic events are entirely repressed in memory for long periods of life?

Persisting flashbulbs that resemble the past

Clearly, we remember details from the traumatic events of our past. We remember the circumstances in which we first learned about shocking public or private tragedies. Take the first major eruption of Mount Saint Helens, which occurred on May 18, 1980. This shocking public event was well known to residents of the northwestern United States.

One field of research that is pertinent to the relationship between emotion and memory, and that most consistently demonstrates high memory performance for negative emotionally arousing events, is research on so-called “flashbulb memories” (see Brown & Kulik, 1977). Since the publication of the original study by Brown and Kulik (1977), numerous studies have shown an impressive concordance in subjects' remembering of shocking national events, such as assassinations, and so forth (see e.g., Bohannon, 1988; Christianson, 1989; Colgrove, 1899; Pillemer, 1984; Rubin & Kozin, 1984; Winograd & Killinger, 1983). Not only does such emotionally shocking news itself appear to be well preserved, but so too do the subjects' memories of the specific detail information associated with the circumstances under which they heard the news. In this chapter, I will discuss how these so-called flashbulb memories compare with memories of other types of emotional events such as personal traumatic events and laboratory-induced emotional events. In this comparison, I will focus on how detailed and persistent these memories are over time. I will also discuss whether a special memory mechanism is involved in remembering flashbulb events as well as other types of emotional events. Before moving on with a discussion of flashbulb memories, I will briefly discuss research concerning memory for negative emotional events. In discussing this research, I use the term “negative emotional events” to refer to scenes or experiences that have very unpleasant features, and that have the potential to evoke strong negative emotional feelings in a victim or a witness.

Thus far this study has attempted an analysis of some of the phenomena associated with the idea of divergent mentalities, and the thrust of my argument has been to suggest that much of the talk of mentalities does nothing to advance and may in certain respects positively impede understanding. Such talk often merely identifies what has to be explained, but without itself providing any adequate explanation: and it may even point away from any hope of one by invoking unverifiable psychological states or processes. In relation to some of the problems that provided part of the original stimulus for the invocation of divergent mentalities, our first step should be to pay due attention to the contexts of communication and in particular to the availability or otherwise of explicit concepts of linguistic categories. The Greek experience is particularly suggestive since we can there study how some such concepts came to be made explicit and the role they played in the development and legitimation of new styles of theorising. In connection, notably, with the Greek contributions to the development of natural scientific inquiry, we have no need to postulate some revolution in mentality, for the key moves consisted, rather, in the development of such new styles of theorising, based on, among other things, a self-conscious use of explicit categories. To understand those developments in turn a variety of factors may be appealed to, but among them the distinctive social and political circumstances of the sixth to fourth centuries b.c. are, it may be argued, of primary importance.

Two of the contexts in which the notion of mentalities has been invoked are (1) in connection with understanding some apparently extravagantly paradoxical or counter-intuitive statements reported in the ethnographic literature as in such famous cases as that of the Nuer belief that ‘twins are birds’, and (2) the general contrast between the scientific and the prescientific. The aim of my first investigation is to tackle some aspects of these issues as they relate to ancient Greek thought with a view to testing how far the concept of distinct mentalities is needed or appropriate. I do not, of course, assume that Greek thought or Greek society bears specific resemblances to those of the Nuer or indeed of any present-day society, though by that I do not mean, either, to claim special, unique status for the Greeks as has been done so often in the past by classicists for whom ‘les Grecs ne sont pas comme les autres’ (cf. the critical remarks of Detienne 1979, ch. 1) let alone by those such as Renan who fantasised over what they thought of as a ‘Greek miracle’ (Renan 1935, pp. 243ff, 1948, p. 393, cf Peyre 1973). But while the notion of mentalities originated as an ethnographic problem, it is, as we have noted, of very general applicability and concerns the historian, the psychologist and the philosopher of science as much as the social anthropologist.