What is memory storage?
The capacity to recall important details and events is at the heart of human existence. People’s identities are merged with who they are, whom they know, and what they have accomplished. The concept of memory has fascinated scientists and psychologists for centuries. Accounts of individuals with superior memory skills, such as those with apparent photographic memories, have further produced interest in the topic of memory. Conversely, the idea of memory loss, especially as a person ages, both frightens and fascinates almost everyone.
Hermann Ebbinghaus is considered to be a pioneer in the experimental investigation of the properties of human memory. After acquiring his doctoral degree, Ebbinghaus set up a laboratory in Berlin, where he formally investigated the concept of memory and forgetting through experimentation. One of his notable concepts was the serial-position effect, wherein the first and last few items in a series are recalled better than the middle items (the primacy effect and the recency effect, respectively). He also described and published his findings on the forgetting curve, in which he described how quickly information is forgotten. Ebbinghaus’s work spurred on the field of experimental psychology, and his concepts continue to be discussed and debated.
Another key historical figure in describing memory was Endel Tulving. Tulving pointed out that up to 1972, most memory research followed in the tradition of Ebbinghaus, focusing on verbal learning tasks concerned with the accuracy of a subject’s performance in remembering personally encountered events (such as the word “bed” in a list). Tulving considered these tasks to be tapping episodic memory.
In 1972, Tulving proposed a distinction between two parallel and partially overlapping memory systems, one for personal experiences and the other for general knowledge of the world. Episodic memory stores personal recollections of episodes or events that one has encountered at a particular time and place. Examples of episodic memories would include remembering such things as one’s first airplane trip, eating cereal for breakfast this morning, and seeing the word August on a list recently learned in an experiment. Semantic memory stores shared factual information about language and the world, but without reference to when it was learned. For example, knowledge that a Boeing 747 is a type of airplane, that cereal is a common breakfast food, and that August has thirty-one days is all stored in semantic memory.
In his 1983 book, Elements of Episodic Memory, Tulving lists twenty-eight differences between episodic and semantic memory. Among them are the source of information (sensation versus comprehension), organization (temporal versus conceptual), emotional content (more important versus less important), vulnerability to forgetting (great versus small), and method of testing in the laboratory (recall of particular episodes versus general knowledge).
Since the early 1970s, memory research has expanded to include tasks designed to uncover the content and organization of information in semantic memory. Some tests assess factual and linguistic knowledge acquired over years of study and experience. Examples include the verbal and quantitative sections of college entrance examinations. A variety of new methods has also been developed to investigate how factual and lexical information is structured and interrelated in long-term memory. For example, subjects have been asked to generate lists of category members, in which the first and most often mentioned instances are interpreted to be the prototypical or best examples available in memory. In a fragment-completion task, the strength of memory traces is indexed by how much of a picture or printed word can be erased from memory and still be identified.
Several semantic memory tasks rely on reaction time as their dependent measure. In a lexical decision task, shorter reaction times to decide whether a string of letters composes a word provide a measure of the item’s strength or current level of activation. Similarly, in a semantic verification task, the time to decide whether a sentence is true (for example, “A tomato is a vegetable”) can be interpreted to indicate the strength of the stated fact or the semantic distance “traveled” between the two named concepts to verify or negate the statement. A robust semantic memory phenomenon is priming, which refers to the activation of associations in a memory network. For example, in a lexical decision experiment, subjects identify nurse as a word more quickly if it is preceded by the semantically related word doctor than by the unrelated word table or the nonword batel. This finding suggests that concepts with similar or shared meanings are stored close to one another in a semantic network, so that accessing one tends to highlight the others.
Memory models based on episodic tasks have focused on the transition of information from acquisition to storage. One of the most renowned models of memory was the multistore model of Richard Atkinson and Richard Shiffrin. This model distinguished between preattentive sensory registers: a limited-capacity, short-term store (STS) and a semantically organized, long-term store (LTS). Of critical interest were the coding processes used to transfer information from STS to LTS. This model gained popularity for some time, but later experimentation identified weaknesses in the model. For example, the famous case study of KF yielded a patient with an extremely limited STS, yet his long-term recall of stories and word lists remained intact. This finding could not be adequately explained by the multistore model of memory. More recently, Fergus Craik and Robert Lockhart proposed a levels-of-analysis model to account for the increasing trace duration of episodic memories. Both of these models were developed to explain the accumulation of memory phenomena from traditional list-learning experiments.
By contrast, semantic memory models have focused exclusively on the organization and retrieval of information already stored in long-term memory (LTM). For example, Lance Rips, Edward Shoben, and Edward Smith proposed a feature-comparison model, in which semantic information is coded as lists of necessary (defining) and descriptive (characteristic) features.
According to this model, subjects verify statements by searching the stored features of the named concepts, looking for matches. Fast reaction times are associated with close matches (resulting in a decision of “true”) or the apparent absence of matches (with a decision of “false”). Their model explains why subjects verify sentences with considerable overlap between concepts (for example, “A robin is a bird”) or none at all (for example, “A robin is a fish”) more quickly than sentences with concepts that have only a few shared features (for example, “A penguin is a bird”). It also accounts for either fast, “false-alarm” errors or slow, correct decisions, when many shared features suggest at first glance that a false statement is true (for example, “A whale is a fish”).
Another model originally proposed by Alan Baddeley and Graham Hitch in 1974 described the components of short-term memory. In this model, Baddeley and Hitch described three major features including a central executive, phonological loop, and visuo-spatial sketch pad. Essentially, the phonological loop is dedicated to rehearsal of verbally based information and the visuo-spatial sketch pad rehearses visual information. Both these components are slaves to the central executive (CE), which functions like the chief executive officer of the system. The CE manages the flow of information to the other components. For example, the CE prioritizes which slave system will be active and incorporates metacognitive (higher order) strategies that facilitate memory. In 2000, Baddeley added a third slave system, the episodic buffer, which serves as a link between short-term and long-term memory. The episodic buffer explains why numbers such as 1492 or 1867 might be easier to remember than other numbers.
An issue of critical importance is how information gets from episodic into semantic memory. Marigold Linton has suggested that as the number of experiences with a particular type of event increases, memories of the specific episodes become confused and eventually cannot be distinguished, but the strength of its generalized trace in semantic memory increases. These contrasting functions were suggested by the results of Linton’s study of her own memory. Every day for six years, she recorded at least two events from her own life, then periodically tested her ability to remember those specific events. She found that memories of unique events, especially those with high emotional content, were often retained intact in episodic memory, whereas repeated events were transformed into generalized, abstracted memories or facts in semantic memory. Overall, her memories of personal events were forgotten at a rate of about 5 percent a year, in a nearly linear fashion.
A class of episodic memories that do not appear to erode over time has been called flashbulb memories. For example, Roger Brown and James Kulik reported that nearly all the adults they interviewed reported vivid personal memories of where they were, what they were doing, how they heard about it, and their own subsequent feelings and actions when they received the news of President John F. Kennedy’s assassination in 1963. Previous generations of Americans have reported similar memories regarding the attack on Pearl Harbor in 1941, whereas other subjects have reported flashbulb memories for the September 11, 2001, terrorist attacks.
The clarity of people’s memories for these and similar important historical and personal episodes has suggested to Ulric Neisser that they constitute a special case of episodic memories that are reflected on, repeated to others, and rehearsed countless times. What is finally remembered is a combination of the original episode, other generic information about the event drawn from semantic memory (for example, recalled facts reported by the news media), and the subtle changes that appear with each repetition. Neisser calls these repisodic memories, and he offers another example in the case of John Dean’s testimony at the Senate’s Watergate hearings in 1973. Dean had practiced the presentations he made to Richard Nixon in the Oval Office, subsequently rehearsed his memories of those conversations many times, and kept a scrapbook of newspaper accounts of the unfolding tale. By comparing Dean’s sworn statements to the tape recordings of the actual White House conversations, Neisser confirmed that Dean’s repisodic memory was essentially correct in retaining the gist of the whole chain of events, even though it did not faithfully report the individual encounters.
Clearly, people’s personal and generic memories are often intertwined, as they recall a compromise between what was and what must have been. From the outset, episodic memories are fashioned by an individual’s preexisting knowledge of the world, which is necessary to make sense of experience. As noted above, semantic memories often reflect generalizations constructed from many similar episodes. As people search their memories, the retrieval process often assumes the characteristics of problem solving. For example, in answering the question “What were you doing at 2:00 p.m. on the third Monday in September five years ago?,” people’s retrieval strategy would depend on some abstract factual information from semantic memory (for example, determining that one was a sophomore in high school that year and would have been in school, in an afternoon biology class). One might then reason that since mid-September is early in the semester, one was probably studying some introductory biological concept such as evolution. That insight then cues an episodic memory of Mr. Brown, the biology teacher, blaming evolution for the football team’s loss in their opening game the previous Saturday night.
Autobiographical memories provide a link to one’s personal past and help one maintain a coherent sense of self. For elderly persons who may be struggling to preserve their self-respect, keeping access to personal memories can be extremely important. Sharan Merriam has reported that providing cues such as old photographs and objects from the past, as part of a technique known as reminiscence therapy, has proved successful with disoriented persons, as it assists them in remembering who they are. Fortunately, even among those patients suffering dense anterograde amnesia (the inability to enter new information into memory), most retain access to autobiographical memories of events that occurred before the onset of amnesia.
Baddeley and Arnold Wilkins have suggested that the episodic-semantic distinction can be meaningfully applied to prospective memory (remembering to perform some act in the future) as well as to retrospective memory (remembering events experienced in the past). An example of a prospective episodic task is remembering to carry out some infrequently performed action on a fixed time schedule, such as recalling to take medication four times a day over the course of a week. In this case, the reference is personal, the organization is temporal, and the occasion of first learning and establishing this intention to act can still be recalled. By contrast, an example of a prospective semantic task is remembering an action sequence of overlearned steps, such as those involved in cooking with a memorized recipe. The reference is cognitive, the organization is cognitive, and the origin of this habitual sequence probably can no longer be recalled.
Douglas Hintzman has noted that authors in various fields had applied labels to capture the essence of the episodic-semantic distinction before Tulving’s landmark article in 1972. In philosophy, Henri Bergson distinguished between pure and habit memory, and Don Locke contrasted personal memory with factual memory. In literature, Arthur Koestler differentiated between picturestrip and abstractive memory. Neurologist Wilder Penfield distinguished between experiential record and concepts. In psychiatry, Ernest Schactel defined autobiographical memory (memory for information and events related to the self from an individual’s past) versus practical memory, and Robert Reiff and Martin Scheerer focused on remembrances versus memoria. Since the 1970s, however, Tulving has been the main standard-bearer of the dichotomy.
Although its heuristic value for classifying memory phenomena and methods has gone relatively unchallenged, Tulving’s claim for episodic and semantic memories as separate systems has engendered considerable debate. Attempts to validate this position have relied on experimental demonstrations of dissociation, or cases in which a variable affects performance in an episodic task differently than it affects performance in a semantic task. Research using functional imaging studies (for example, functional magnetic resonance imaging, or fMRI) have, in general, not supported the distinction proposed by Tulving. These studies have suggested that multiple regions of the brain work together in the formation of new memories. Therefore, the process of memory consolidation is much more complex than originally hypothesized.
A notable outcome of Tulving’s original argument was to shift attention in memory research away from episodic list learning and toward the structure of long-term semantic knowledge. However, there has been a growing interest in personal episodic, or autobiographical, memory. These are precisely the kinds of recollections that Ebbinghaus sought to exclude from the study of memory, considering them too personal and difficult to verify. Any comprehensive theory of memory, however, will need to account for the full range of interdependent memories, from highly personal accounts to abstract facts.
Injury or damage to specific brain regions have been informative about the process of memory. Specifically, disruption to a person’s stream of consciousness has predictive value for the recovery of a patient following head trauma. Loss of memory before the event is called retrograde amnesia, whereas loss of memory of events following the event is called anterograde amnesia. Typically, the longer the time period of amnesia after the accident, the more likely there will be permanent impairment of brain functioning. Disruption to an individual’s memory, especially short-term memory, is a common outcome of many neurological and cardiovascular illnesses or injuries. This reflects the brain’s vulnerability to memory loss following trauma.
An interesting although controversial topic relates to what is commonly described as photographic memory. Also called eidetic memory, persons with this rare ability are said to have an extraordinary capacity to recall precise details. Classic experiments determine that these individuals can study a detailed picture for thirty seconds and then maintain a complete visual image of the picture. It is believed that historical figures including Napoleon and Theodore Roosevelt may have had this capacity. In rare cases, extraordinary memory capacity has been identified in individuals with developmental disabilities such as autism.
Baddeley, Alan D. Human Memory: Theory and Practice. Rev. ed. Hove: Psychology, 2005. Print.
Bauer, Patricia J. Remembering the Times of Our Lives: Memory in Infancy and Beyond. New York: Psychology, 2013. Print.
Carruthers, Mary J. The Book of Memory: A Study of Memory in Medieval Culture. Rpt. New York: Cambridge UP, 1993. Print.
Emilien, Gerard, et al. Memory: Neuropsychological, Imaging, and Psychopharmacological Perspectives. New York: Psychology, 2004. Print.
Kahana, Michael Jacob. Foundations of Human Memory. New York: Oxford UP, 2012. Print.
Neisser, Ulric, and Ira E. Hyman, eds. Memory Observed: Remembering in Natural Contexts. 2nd ed. New York: Worth, 2000. Print.
Taylor, Annette Kujawski, ed. Encyclopedia of Human Memory. 3 vols. Santa Barbara: ABC-CLIO, 2013. Print.
Tulving, Endel. Elements of Episodic Memory. London: Oxford UP, 1983. Print.
Tulving, Endel. “How Many Memory Systems Are There?” American Psychologist 40.4 (1985): 385–98. Print.
Tulving, Endel, and Fergus I. M. Craik, eds. The Oxford Handbook of Memory. New York: Oxford UP, 2000. Print.