Although long-term memory was treated as a single type of memory in these early psychological accounts, neuroscientists soon found that memory for different kinds of information could be differentially affected by brain damage. In an attempt to control severe epileptic seizures in a patient designated "H.M.,"
William Scoville removed much of his medial temporal lobe.4 The surgery was successful in relieving H.M.'s epilepsy, but left him severely amnesic (Scoville & Milner, 1957). In particular, after the surgery he was not able to learn new information or remember events in his life since the surgery (anterograde amnesia). Moreover, he experienced graded loss of memory for several years of his life preceding the surgery (retrograde amnesia). Nonetheless, H.M.'s short-term memory was unimpaired—if not distracted he could retain information for several minutes. One result of the line of research that began with H.M., to which I will return below, was the hypothesis that the hippocampus plays a critical role in memory. But equally important was that behavioral experiments with H.M. showed that he could acquire new skills even though he had no memory of learning them (Corkin, 1968). This suggested that skill memory was a distinct type of long-term memory.
Shortly thereafter, Endel Tulving (1972) proposed a further fractionation of long-term memory in which he distinguished "two parallel and partially overlapping information processing systems" (p. 401). One, episodic memory, is concerned with events or episodes in a person's own life, specifying information about the time and place of their occurrence and permitting the person to "travel back" to reexperience them. The other, semantic memory, is concerned with information that typically is retrieved independently of recalling the time and place in which it was acquired (e.g., word meanings, general knowledge, scientific facts). One difference between these types of memory is that episodic memory closely involves the self, whereas semantic memory seems more removed. That is, remembering a significant event in your own life involves you in a fundamentally different way than does your knowledge about Plato. James captured this aspect of episodic memory: "Memory requires more than the mere dating of a fact in the past. It must be dated in my past ... I must think that I directly experienced its occurrence" (James, 1890/1950, p. 612).
Tulving introduced the term autonoetic awareness for the access one has to episodic memories. Autonoetic awareness involves being able to move around in the temporal dimension of one's life (Tulving often speaks of episodic memory as involving "mental time travel") and reexperiencing events that one has previously experienced (a capacity that, like memory more generally, is error-prone: people sometimes have compelling memories of events that never happened). Tulving quoted James to capture the affective aspect of autonoetic awareness—its "feeling of warmth and intimacy." He contrasts such remembering with knowing, which he regards as characteristic of semantic memory. One may know about a number of events in one's life (e.g., that one was born in a particular place) without reexperiencing the event.
4 Scoville thought he had removed H.M.'s hippocampus, but later neuroimaging studies revealed that the lesion spared much of the hippocampus proper, although it subsequently atrophied as a result of loss of its normal inputs from surrounding cortical structures (Corkin, Amaral, González, Johnson, et al., 1997).
Such recollections count as semantic, not episodic, memories. Tulving has articulated a number of additional distinctions between these two kinds of memory. For example, episodic memory is oriented towards the past, whereas semantic memory applies information learned in the past to the present. Episodic memory, moreover, is evolutionarily more recent—he contends that there is no evidence of episodic memory in any species other than humans—and it develops later in ontogeny (Tulving, 1999a).
In his initial formulation, Tulving proposed that different types of tasks measured episodic and semantic memory performance. His tests of episodic memory involved recall and recognition of recently studied events, whereas semantic memory was tested by such tasks as generating words from a fragment or from a definition, identifying words from brief tachistoscopic displays, and making lexical decisions (i.e., presented with a series of letter strings, deciding whether each constitutes an English word). In one experiment, Tulv-ing and his collaborators tested episodic memory by measuring recognition of words presented in a list-learning task. Either one hour or seven days later, they tested semantic memory by asking the same subjects to fill in the missing letters in incomplete words, such as _o_ma__c. Although completing such a word fragment can be difficult, it was rendered much easier if the target word (here, "dogmatic") was on the list used earlier in the episodic-memory task. Success in fragment completion remained roughly constant over the one-week interval, whereas recognition performance declined dramatically. Such dissociation between results is often construed as evidence that the psychological processes are independent. In a further analysis, it was established that the probability that subjects could complete a given fragment was stochastically independent of the probability that they could recognize it as having been on the study list (Tulving, Schacter, & Stark, 1982), providing additional evidence that different processes were involved. Subsequently Tulving emphasized that the difference in the kind of information remembered (general facts versus personally experienced events) was more fundamental.
Although Tulving advanced the distinction between episodic and semantic memory on behavioral grounds, he found powerful confirmation in studies of the patient "K.C.," who became profoundly amnesic following a closed-head injury in 1981. Whereas H.M. suffered substantial loss of both semantic and episodic memory, K.C. lost all episodic memory (retrograde and anterograde), while retaining much of his semantic memory:
The outstanding fact about K.C.'s mental make-up is his utter inability to remember any events, circumstances, or situations from his own life. His episodic amnesia covers his whole life, from birth to the present. The only exception is the experiences that, at any time, he has had in the last minute or two. It does not matter how much and how specific information is given to him about any particular event from further back in the past, how memorable the event is by ordinary standards, how long its duration, or how many times he has been asked about it before. He always denies any recollection and does not even acknowledge any sense of familiarity with the event (Tulving, 2002, p. 14).
K.C. nonetheless retains his factual knowledge, including knowledge of facts about himself: he can report the date of his birth, the names of many of the schools he attended, the make and color of a car he had owned, etc. In reporting these he relies on semantic memory—he does not remember experiencing any of these but simply knows about them. Moreover, with extensive training, K.C. has been able to learn both new skills and new semantic information. He thus presents a dissociation between episodic and semantic memories.
An even stronger dissociation of episodic and semantic memory was reported by Faraneh Vargha-Khadem et al. (1997). This involved three patients who suffered bilateral damage to the hippocampus at ages ranging from birth to nine years of age. All suffered severe impairments in episodic memory, especially for events of their daily lives (where they had left an item, what television show they had watched, what conversations they had had). Nonetheless, all have been successful in school. They have developed language skills commensurate with their age and are able to learn and reason about information that involves semantic memory (Squire & Zola, 1998).
Tulving's distinction between episodic and semantic memories was not the only tax-onomic distinction developed during this period. Neal Cohen and Larry Squire (1980) approached the fractionation of long-term memory in a way that emphasized what episodic and semantic memories have in common. In particular, people can report semantic knowledge and episodic memories linguistically. But there are other memories people can acquire with experience but cannot report linguistically, such as the ability to ride a bicycle or solve crossword puzzles (these are the sorts of skills H.M. and K.C. could both acquire). Typically, people cannot explain in words how they perform these skills. Philosopher Gilbert Ryle (1949) characterized the distinction between skill knowledge and semantic knowledge as knowing how versus knowing that; Cohen and Squire characterized it in terms of procedural versus declarative information. As an individual could report both semantic and episodic memories in language, both counted as declarative for Cohen and Squire.5
5 An important difference between Cohen and Squire's approach and Tulving's is that Cohen and Squire were guided in part by work on animal models. At first it may seem surprising to consider declarative memory in organisms who cannot report their memories in language. But animal investigators have found ingenious ways to detect declarative memories in animals. One approach involves testing whether the animal recognizes a previously presented stimulus. Researchers, for example, can present a monkey with a stimulus, remove it, and then present two further stimuli, one of which matches that presented before. The monkey can then be trained to select the new stimulus (this procedure is referred to as delayed nonmatching to sample). Researchers have found that hippocampal damage in monkeys impairs ability on such tasks (Zola, Squire, Teng, Stefanacci, & Clark, 2000).
Skills Priming Simple Non-
Episodic habits classical associative conditioning learning
Emotional Skeletal responses musculature
Skills Priming Simple Non-
Episodic habits classical associative conditioning learning
Emotional Skeletal responses musculature
Figure 2.2 Squire and Knowlton's (1995) taxonomy of memory systems.
An important aspect of declarative memories is that they involve a representation of external objects and events and accordingly can be evaluated in terms of whether they provide a true or false representation of those objects or events. Procedures, on the other hand, may be appropriate or inappropriate, but are not judged as true or false. Moreover, although procedures generally take practice to acquire, declarative memories can sometimes be acquired in a single exposure. Finally, declarative memories can be flexibly expressed in multiple forms (verbal reports, problem solving), whereas procedures are often far less flexible and transferable to other expressions.
Once researchers extended their attention beyond declarative memory, they came to realize there are subtypes of nondeclarative memory and that some (e.g., skills and habits) more clearly relate to the procedural notion than others. Figure 2.2, adapted from Squire and Knowlton (1995), is representative of the taxonomies proposed. As indicated there, the original dichotomy has been reconceptualized as a contrast between explicit and implicit memory, often characterized in terms of whether conscious awareness is involved. Daniel Schacter (1987), for example stated: "Implicit memory is revealed when previous experiences facilitate performance on a task that does not require conscious or international recollection of these experiences" (p. 501). This raises the question as to whether there is any cohesiveness to the category of implicit memory (Willingham & Preuss, 1995). Schacter's statement emphasizes a subtype of implicit memory called priming. In priming, exposure to a stimulus that is not explicitly remembered may, nonetheless, bias responses in later situations. It is often assessed using a variation of the fragment completion task, the stem completion task, in which a subject will be asked to complete a stem such as sal___with the first word that comes to mind. If the word salary was on a list of words that subjects had previously read, then even if they could not remember or recognize that salary had been on the list, they would still be more likely to complete the stem with salary than if it had not been on the list. Priming exhibits a different behavioral profile than explicit memory. For example, factors such as typeface are more likely to affect priming than explicit recall (Roediger & McDermott, 1993). Moreover, priming is often spared in patients with amnesia (Squire, 1987). Priming, however, appears not to be a single phenomenon. Researchers have found dissociations between perceptual priming, which is linked to a specific sensory modality (Church & Schacter, 1994) and conceptual priming, which is revealed in tasks requiring semantic analysis such as generating exemplars for a category (Hamman, 1990). Moreover, forms of priming can be dissociated from skill learning, which is preserved in Alzheimer patients who exhibit deficits in priming tasks (D. P. Salmon & Butters, 1995).
Closely linked to the notion of implicit memory is that of implicit learning, a concept first developed by Arthur Reber (1967) to characterize the ability of people to learn complex skills without conscious awareness. Reber developed artificial grammars in the form of finite state automata (often referred to as Reber grammars) used to create strings of letters. Subjects were presented with some number of strings generated by the grammar (e.g., TSSXS and PTTSS) and then asked to determine whether additional strings presented to them (e.g., TSXXS) followed the same rules. Even though subjects were not able to figure out the rules, they were significantly above chance in distinguishing strings that fit the grammar from strings that did not.
Was this article helpful?
If Pregnancy Is Something That Frightens You, It's Time To Convert Your Fear Into Joy. Ready To Give Birth To A Child? Is The New Status Hitting Your State Of Mind? Are You Still Scared To Undergo All The Pain That Your Best Friend Underwent Just A Few Days Back? Not Convinced With The Answers Given By The Experts?