Psychology: Explicit and Implicit Memory

Explicit and Implicit Memory

We’ve now considered several ways that explicit memory might be subdivided—into episodic memory and semantic, and then with each of those categories potentially divided further. But what about implicit memory? As we’ll see, this memory provides an entirely different means through which we’re influenced by past experience, and it is distinct from explicit memory in its functioning and in its biological basis.

DISTINGUISHING IMPLICIT FROM EXPLICIT MEMORY

Implicit memories are distinguishable from explicit memories in many ways. Perhaps the clearest evidence, however, comes from the study of the memory disruption caused by brain damage. In general, this disruption is referred to as amnesia. Earlier, we mentioned one type of amnesia: retrograde amnesia—a loss of memories for events that took place before the brain injury that caused the amnesia. In other cases, though, brain damage produces anterograde amnesia—an apparent inability to form new memories (see Figure 8.9).

In general, anterograde amnesia is caused by damage to certain sites in the temporal cortex—specifically, in the hippocampus and nearby subcortical regions. In some cases, this damage is the result of illness—especially if the illness causes encephalitis, an inflammation in the brain tissue. In other cases, the damage is caused by stroke or physical trauma. One of the most common causes, though, is a type of malnutrition associated with chronic alcoholism; in this case, the amnesia is a central symptom of the illness called Korsakoff ’s syndrome.

One of the most carefully studied cases of anterograde amnesia, however, had an entirely different cause: A patient known as H.M. suffered from severe epilepsy. When all other treatments failed, the physicians tried (in the late 1950s) to treat H.M.’s disease with a neurosurgical procedure that deliberately removed most of his hippocampus, amygdala, and a considerable amount of nearby tissue (Figure 8.24). The procedure was, in a very narrow sense, a success: It did control his epilepsy. But the surgery also had a tragic and unanticipated side effect. Across the 55 years he lived after his surgery, H.M. seemed incapable of adding new information to his long-term memory. As his obituary put it: “each time he met a friend, each time he ate a meal, each time he walked in the woods, it was as if for the first time” (Carey, 2008). He remembered none of the episodes in his life after the surgery; he was entirely unable to recognize people he’d first met after the surgery—even if he saw them day after day (Milner, 1966, 1970; also see O’Kane, Kensinger, & Corkin, 2004; Skotko et al., 2004).

Amnesia had devastating effects on H.M.’s life—including some effects that we might not think of as involving memory. For example, H.M. had an uncle he liked very much. When first told that his uncle had died, he was deeply distressed, but then he for-got all about this sad news. Some time later, he asked again when his uncle would come to visit, and he was told again of his uncle’s death. His grief was as intense as before; indeed, each time he heard this sad news, he was hearing it for the first time—with all the shock and pain (Corkin, 1984; Hilts, 1995; Marslen-Wilson & Teuber, 1975; Milner, 1966; Milner, Corkin, & Teuber, 1968).

Crucially, and despite these remarkable problems, patients with anterograde amnesia— including H.M.—can acquire certain types of new memories which can be revealed with specialized testing. In some studies, for example, patients with anterograde amnesia have been given practice, day after day, in finding the correct path through a maze. Each time they’re shown the maze, the patients insist they’ve never seen it before; this is simply a con-firmation of their amnesia. Even so, they get faster and faster in solving the maze—and so apparently they do retain some information from each practice session.

Likewise, in another study, patients with Korsakoff ’s syndrome heard a series of brief melodies (Johnson, Kim, & Risse, 1985). A short time later, they listened to a new series and were told that some of the tunes in the second batch were repeats from the earlier presentation. As expected, these amnesic patients were completely unable to tell which tunes were the repeats and which were new; indeed, their memory responses were close to random. Remarkably, though, when asked which melodies they preferred, the patients uniformly preferred the familiar ones. The patients had no (explicit) memory for these tunes, but a memory did emerge with indirect testing—and emerged, in this case, as a preference.

In important ways, therefore, these patients can’t remember their experiences. If we ask them directly about the past, they recall nothing. If we ask them which mazes they have solved before and which are novel, they can only guess. Thus it seems clear that these patients have no conscious recollection, no explicit memory, for the events in their lives. Still, we can find ways in which the patients’ current skills and behaviors are shaped by their experiences—and so, apparently, the experiences have left some record, some residual imprint, in these patients. This lasting imprint, a demonstrable impact of the past, is what psychologists call implicit memory—an unnoticed “leftover” from life events that changes how someone now acts and thinks (Donaldson, Peterson, & Buckner, 2001; Fazio & Olson, 2003; Humphreys et al., 2003; Kinoshita, 2001; Yonelinas, 2002).

PROCEDURAL MEMORY

What exactly is implicit memory? In what circumstances does it influence us? And can implicit memory be demonstrated in people without amnesia, people whose brains are healthy and intact? To answer these questions, we need to distinguish different types of implicit memory, because each type influences us in its own way.

Some cases of implicit memory involve procedural knowledge rather than declarative knowledge. Procedural knowledge is knowinghow—knowing how to ridea bicycle, for example, or how to use chopsticks. Declarative knowledge, in contrast, is represented in explicit memory, not implicit, and it’s knowingthat: knowing that there are three outs in an inning, that automobiles run on gasoline, or that you woke up late this morning.

The earlier example we mentioned in our discussion of amnesia—that of patients learning how to get through a maze—involves procedural memory, and other examples are easy to find. In some procedures, patients have been shown a complex shape and asked to trace the outline of the shape with a stylus. What made this task difficult was that the patients couldn’t see the shape directly; they could see it (and the stylus) only by looking into a mirror (Figure 8.25). This task is moderately difficult—but the patients got better with practice, all the while insisting that each try at the task was their very first time.

What about people who don’t have amnesia—people with normal brains? In some studies, research participants are given four buttons and told to press button 1 if light 1 comes on, button 2 for light 2, and so on (Figure 8.26). The lights are then turned on in rapid succession, and participants do their best to keep up. As it turns out, the lights are turned on in a repetitive sequence—perhaps always 1-3-4-2-1-4-1-3-4-2-1-4-1-3-4-2-1-4. Participants seem to learn this sequence; and so, with a bit of practice, can respond more quickly if the lights follow this sequence than they can if the sequence is random. But when asked whether the sequence was random or not, participants are clueless. Thus, they seem to have procedural knowledge that allows them to respond more quickly to the patterned lights, but they don’t have declarative knowledge—the same distinction we observe in patients with amnesia (Gazzaniga et al., 2009).

PRIMING EFFECTS

Procedural memories are typically concerned with behaviors—our actions and our skills. Other types of implicit memory, in contrast, influence our perceptions and our thoughts. Consider, for example, demonstrations of priming. Participants in one study were shown a number of words. Later, they were given a second task in which they sim-ply had to identify words flashed briefly on a computer screen. Participants had no idea that many of the words in this second task were taken from the earlier list, but they still showed a pattern known as repetition priming: Words that had been on the original list were identified more readily than words that had not. This priming was observed even for words that the participants failed to recognize as familiar in a standard recognition task. Thus, the participants had no explicit memory for having seen these words, but they did have an implicit memory that showed up as priming. In other words, they were being influenced by a memory they didn’t realize they had (Jacoby, 1983; Jacoby & Witherspoon, 1982).

Other procedures, with different tasks, show a similar pattern. In fragment-completiontasks, for example, participants are shown partial words (such as C_O_O_I_E) andasked to complete them to form actual words (CROCODILE). Success in this task is much more likely if the target word was encountered recently; this advantage is observed even when participants have no conscious recollection of the previous encounter (Graf & Mandler, 1984; Jacoby & Dallas, 1981; Tulving, Schacter, & Stark, 1982).

In another experiment, participants were asked to read sentences that were pre-sented to them upside down. A year later, participants returned to the lab; there, they were shown a series of sentences and asked which ones they’d seen in their first visit to the lab and which ones were novel. Not surprisingly, after this long delay, participants couldn’t tell which sentences they’d seen before. Still, when they were asked once again to read sentences presented upside down, they were faster with the sentences they’d seen before than they were with novel sentences—a case of priming that lasted across a full 12 months (Kolers & Roediger, 1984).

In each of these cases, it seems that an encounter with a stimulus leaves us better prepared for that stimulus the next time we meet it. This preparation can then influence us in many ways, quite independently of whether we can recall the earlier encounter with that stimulus. To illustrate how far this pattern can reach, consider the so-called illusion of truth. In the relevant studies, participants hear a series of statements like “The average person in Switzerland eats about 25 pounds of cheese each year,” or “Henry Ford forgot to put a reverse gear in his first automobile.”* Participants’ task is to say how interesting each of these statements is. Later on, the same participants are presented with some more sentences but now have to rate the credibility of each one on a scale from “certainly true” to “certainly false.” Needless to say, some of the sentences in this “truth test” are repeats from the earlier presentation; the question for us is how the judgments of sentence credibility are influenced by the earlier exposure.

The result of these studies is a propagandist’s (or advertiser’s) dream: Sentences heard before are more likely to be accepted as true, so that in essence familiarity increases credibility (Begg, Anas, & Farinacci, 1992). To make matters worse, the effect emerges even when participants are warned not to believe the sentences in the first list. That is, sentences plainly identified as false when they’re first heard still create the illu-sion of truth, so that these sentences are subsequently judged to be more credible than sentences never heard before.

How could this be? Bear in mind that the participants in these procedures are shown a lot of sentences and that there’s a delay between the first task (judging how interest-ing the sentences are) and the second (judging credibility). These steps make it difficult for participants to keep track of the sentences they hear; in other words, these steps work against explicit memory. As a result, participants have a hard time recalling which of the sentences in the truth test they encountered on the first list; so it doesn’t help them to know that the sentences on that first list were all false. Thus, with no conscious memory of the earlier encounter, participants have no way to protect themselves from the illusion.

OTHER FORMS  OF  IMPLICIT MEMORY

We’ve now mentioned two forms of implicit memory—priming effects and procedural memory—but there are other forms as well. One plausible addition to this list is perceptual learning—the learning that you need to do whenever you “recalibrate” yourperceptual systems. As an example, think of what happens when someone gets new eyeglasses, perhaps with a stronger prescription than they’ve had before. Across the next few days, he needs to “adjust” to the glasses—changing (among other things) how he interprets the degree of tension in his eye muscles as a cue to distance. This is surely a form of learning—and so places new information, or perhaps new skills, in memory. But it’s learning that happens completely outside of awareness—and so involves implicit memory, not explicit.

A different example involves cases including the learn-ing called classical conditioning. This learning, too, creates new knowledge—knowledge about what follows what in the world—but can be done without conscious awareness. Indeed, classical conditioning can take place even if an organism is fully anesthetized during the learning (e.g., Cahoon, 2008).

These and other examples demonstrate the enormous breadth of implicit memory. We rely on explicit memory in many circumstances, and we’re guided to an enormous extent by our conscious recollection of the past. But the reach of implicit memory may be even larger—so that in many situations, we’re shaped in ways we do not notice by past experiences that we cannot recall.