Home Curriculum Vitae Publications Conference Reports Forthcoming Extramural Colloquia Expert Testimony Teaching Healthcare The Human Ecology of Memory Research Archive Publications and Reports Rants

Interactions Between

Implicit and Explicit Memory


John F. Kihlstrom

University of California, Berkeley


Episodic memory, recording an individual's own past events and experiences, can be expressed both explicitly and implicitly (Schacter, 1987). Explicit memory means the conscious recollection of the past. We usually think of it as reflected in our ability to recall and recognize our past actions and experiences. Implicit memory refers to the effect that past events and experiences have on our current experience, thought, and action, even if these ongoing activities do not involve, or require, conscious recollection of the past. Implicit memory is exemplified by priming effects, which occur when the prior occurrence of an event facilitates or inhibits the processing of a subsequent, related event. Thus, a subject who has previously studied the word coelacanth will be more likely to succeed when asked to complete the fragment c_e_a_a_t_ with a legal English word, or when asked to name the ancient fish caught in the Indian Ocean.

In principle, explicit and implicit memory differ in terms of task demands: explicit memory tasks requires conscious recollection, while implicit memory tasks do not. But more important, explicit and implicit memory are dissociable in fact. Priming and other manifestations of implicit memory are spared in amnesic patients, who show gross impairments in explicit memory. Furthermore, studies of nonamnesic subjects, show that different variables affect performance on explicit and implicit memory tasks. As we have heard earlier in this program, the fact that explicit and implicit memory are dissociable has led some theorists to conclude that they are based on separate and independent memory systems. So, for example, it might be that explicit memory is mediated by the hippocampus and other structures which make up the medial temporal lobe memory system, while implicit memory is mediated by cortical perceptual representation systems.

Well, maybe. I have no doubt that various parts of the brain play different roles in explicit and implicit memory, but I worry that implicit memory has been defined too narrowly, and that the independence of explicit and implicit memory has been asserted too strongly.

First, consider that from the beginning, the study of implicit memory has been dominated by a single paradigm, priming, and only a very specific form of priming, repetition priming, at that. In repetition priming, the prime and target are identical, or the target is some kind of portion of the prime. Thus, in perceptual identification the prime is presented again, albeit very briefly or degraded by a mask or noisy background, as the target. In picture fragment completion, fragments of the prime are presented as the target. But there are other kinds of priming besides repetition priming. For example, long ago I demonstrated that posthypnotic amnesia impaired free recall of a word list studied during hypnosis, while leaving intact the ability of the amnesic subject to generate list items as free associates or as category instances (Kihlstrom, 1980). More recently, Jennifer Dorfman and I have confirmed this observation in more adequately designed experiments (Dorfman & Kihlstrom, 1994). Our subjects, selected for very high levels of hypnotizability, memorized a list of 32 items, after which they received a conventional suggestion for posthypnotic amnesia. After termination of hypnosis, we tested half the items explicitly, and the other half implicitly. For the explicit test, we presented the subjects with a cue word selected to have a high probability of eliciting one of the items as a free associate, and asked them to produce an associated word from the study list. In the implicit test, we presented subjects with the same sorts of cues, and asked them simply to tell us the first word that came to mind. A control group memorized the list in the normal waking state, without any suggestions of amnesia. The result was a double dissociation: posthypnotic amnesia impaired performance on the explicit test, but enhanced performance on the implicit test. The enhanced performance on the implicit test is, of course, an example of priming. But this priming is semantic in nature, and cannot be mediated by purely perceptual analyses performed by a perceptual memory system.

This is not the only case in which semantic priming is dissociated from explicit memory. There are others, both in normal subjects and in cases of amnesia secondary to brain damage. But the nature of this dissociation has not been subject to very much experimental investigation. The debate over implicit memory has been almost completely dominated by studies of repetition effects observed perceptual identification, lexical decision, stem and fragment completion, picture fragment naming, and judgments about objects, and I worry that our theories of implicit memory may have been distorted accordingly.

The worry is compounded with we consider that priming is not the only way that the past can implicitly influence the present. There are lots of other implicit memory tasks, and we know very little about them or how they relate to the repetition priming which we have studied so much. Even familiar verbal-learning paradigms can be reconstrued in terms of implicit memory. So, for example, in an experiment in which the subject studies list AB, then studies list AC, and then must remember list AB, the retroactive interference of AC on AB classifies is an implicit memory effect. To be sure, the subject is being asked to consciously remember something, but what s/he's being asked to remember is AB, not AC, and it's implicit memory for AC which is getting in the way. We know almost nothing about the conditions under which indirect memory as displayed in interference and savings paradigms relates to implicit memory as displayed in priming paradigms. We know very little about how implicit memory is related to implicit learning, and procedural learning in general. And we probably ought to find out.

Turning to what we do know, one piece of evidence for the dissociation between explicit and implicit memory is that explicit memory is affected by processing activity at the time of encoding, but implicit memory is not. Thus, for example, Jacoby & Dallas (1981) had subjects perform "shallow" physical and "deep" semantic processing tasks on a list of items, and found that level of processing affected explicit recognition but not implicit perceptual identification. This is the canonical result in implicit memory experiments, and it has played a major role in theory, but the first thing you should note is that almost all of these experiments have employed repetition priming tasks of the sort that can be mediated by purely perceptual processing at the time of encoding. It should surprise nobody if deep, semantic processing plays a relatively small role in performance on such a task. Roediger and his colleagues made this point somewhat differently, when they pointed out that most explicit memory tasks are conceptually driven, involving elaboration, organization, and meaning analysis, while most implicit memory tasks are data-driven, involving no more than the analysis of perceptual structure (e.g., Roediger & McDermott, 1990). In a review by Brown and Mitchell (1994) of the effects of levels of processing on implicit memory, only 12 studies involved conceptually driven, semantic, implicit memory tests. Despite demonstration by Roediger's group of dissociations between perceptual and semantic implicit memory, which should have told us something about the extent to which we can generalize from the perceptual case, research on implicit memory continues to be tightly focused on repetition priming. I think this is a mistake.

Even in the repetition case, though, the dissociation observed in levels-of-processing experiments isn't secure. For example, Challis and Brodbeck (1992) surveyed the literature and found that in 33 of 35 cases, priming on perceptual identification or stem- or fragment completion was numerically greater in the semantic than in the physical condition; and they confirmed this observation in experiments in which level of processing was manipulated between subjects, or within subjects in lists blocked by condition; when the manipulation is within subjects in mixed lists, the effect of level of processing is not significant, but that's probably a problem of power. Level of processing may affect explicit memory more than it affects implicit memory, but there is a levels effect, small to moderate in magnitude, and it has to be explained. One possibility is that during the perceptual identification test, subjects recognize some targets as items from the study phase. If so, they may consciously retrieve other study items in order to prepare themselves for what they might see next -- what Jerome Bruner might call a strategic increase in perceptual readiness. This is one way in which explicit and implicit memory can interact.

The other side of the interaction is revealed by the effects of implicit on explicit memory. That such effects occur might offer a solution to a longstanding puzzle about amnesia. We know that amnesia impairs recall and spares priming, but amnesics sometimes show significant levels of recognition as well. For example, Huppert and Piercy (1977) found that even H.M. can show normal levels of recognition memory under some conditions. Hirst, Johnson, and their colleagues (Hirst, Johnson, Kim, Phelps, Risse, & Volpe, 1986) tested alcoholic and nonalcoholic amnesic patients on recognition of a list of 40 words, five items in each of eight categories. Compared to controls, recall was severely impaired, but recognition was completely spared.

These findings are interesting, because under many theories of amnesia, this shouldn't happen. If amnesia entails a dissociation between episodic and semantic memory, the problem is that both recall and recognition are episodic. If amnesia entails a dissociation between declarative and procedural (or nondeclarative) memory, both are declarative. If amnesia entails a dissociation between explicit and implicit memory, both are explicit. Of course, recognition is in some sense easier than recall, and thus should yield more memory, especially in patients who are not completely amnesic, but I think that the answer is more interesting than that. I suspect that, just as subjects can use explicit memory to enhance their readiness for an implicit memory task, they can also capitalize strategically on implicit memory to enhance their performance on an explicit memory task.

This view is based on the two-process theory of recognition proposed by George Mandler in 1980. Mandler began with the insight that recognition is a judgment of prior occurrence, and he proposed that this judgment is based on two quite different processes: retrieval, when an event corresponding to the target is consciously recollected, and familiarity, when the target just "rings a bell". Retrieval is based on the recovery of an episodic memory trace, including information about the spatiotemporal context in which the event occurred (and, I might add, about the role of the self as the agent or experiencer of the event; Kihlstrom, 1996). Familiarity, by contrast, is based on the activation by the target of pre-existing memory structures, independent of the episodic context in which the trace was encoded. These two recognition processes, in turn, are related to explicit and implicit memory. Retrieval obviously constitutes explicit memory, involving as it does the conscious recollection of an event. On the other hand, familiarity has the character of priming. The implication is that if priming is spared in amnesia, then we ought to see preserved recognition as well, provided that the amnesics are encouraged to capitalize on the perceptual (and conceptual) salience which is part and parcel of what priming is all about.

Let me illustrate what I mean with a study of recognition in posthypnotic amnesia. In this study, hypnotized subjects memorized a list of 16 items, and then received a suggestion for posthypnotic amnesia. After the termination of hypnosis, recognition was tested with a four-point confidence scale. Such a scale yields three criteria for recognition: strict, counting only the 4s; moderate, counting the 3s as well; and liberal, counting even 2s. When we plot hits and false alarms, we see that hits increase appreciably as we loosen the criterion for recognition; false alarms also increase, but not at the same rate, so that there is a genuine improvement in recognition memory as indexed by d-prime. These subjects aren't merely lowering their criterion for saying "yes".. Rather, I think that they are adding a criterion of familiarity to that of retrieval, as they make judgments about their personal past.

We can see this process in other amnesias as well. Consider an experiment by Jennifer Dorfman and myself, on the retrograde amnesia associated with electroconvulsive therapy (Dorfman, Kihlstrom, Cork, & Misiaszek, 1995). In this study, a series of patients receiving ECT for depression studied a wordlist immediately before one of their treatments. Once they regained consciousness in the recovery room, they were tested with three-letter stems of list items and matched controls. As expected, these patients showed a dissociation between explicit and implicit memory: a profound deficit on stem-cued recall, but strong priming on a test of stem completion. Then the subjects received a test of recognition: for half the items, they were instructed to adopt a very conservative criterion, saying yes only if they were fairly sure that the item was on the study list; for the other half, they were encouraged to adopt a more liberal criterion, saying yes if the item rang a bell, even if they were uncertain. When the subjects shifted criterion, hits went up but false alarms stayed constant, resulting in a significant increase in d-prime. This genuine improvement in recognition memory occurred, we think, because the subjects were encouraged to capitalize on the experience of perceptual and conceptual fluency which is the phenomenal accompaniment to priming.

Recently, Squire and Reber have reported a failure to replicate these findings, so let me quickly say a couple of things in our defense. In the first place, their first two experiments didn't replicate our procedures. We deliberately delayed the recognition test until after the priming test, while they chose to place the recognition test first. The failure to replicate our procedures makes it impossible to interpret their failure to replicate our results. Their third experiment replicated our procedures, but unaccountably failed to produce priming on the stem completion test. Recognition by familiarity assumes that priming will be at least relatively spared, so that experiment is indecisive as well. So our conclusion, that priming can contribute to performance on a recognition test, is actually unchallenged by these experiments.

Further evidence on the contribution of priming to recognition is provided by recent experiments by Michael Kim and myself on recollective experience in episodic recognition (Kihlstrom, Kim, & Dabady, 1996; Kim & Kihlstrom, 1997). Endel Tulving (1985) and John Gardiner (1988) have distinguished between two such experiences: remembering, the concrete awareness of oneself in the past; and knowing, one's abstract knowledge of the past. For Tulving, the remember-know distinction maps onto his distinction between episodic and semantic memory. Gardiner, however, has argued that it maps more closely onto the distinction between retrieval and familiarity, or between explicit and implicit memory. Kim and I, however, had the sense that there are at least three different varieties of recollective experience -- or, if you will, three different memory "qualia": remembering involves the conscious recollection of some past event, as an explicit expression of episodic memory; knowing is abstract knowledge of that event, as an item in semantic memory; and feeling is the intuition that an event occurred in the past, a product of the salience which accompanies implicit expressions of episodic memory. Any of these can form the basis for a judgment of prior occurrence. And frankly, I suspect that there is a fourth basis as well: believing, in which we surmise that an event occurred in the absence of any recollective experience at all. You might think that believing is more like judgment than memory; but if you follow Mandler, not to mention Bartlett, you understand that all acts of remembering are ultimately acts of judgment. Remembering is solving the problem of making sense of the past.

Anyway, Kim and I separated remembering, knowing, and feeling in a series of experiments closely patterned on those reported by Tulving and by Gardiner. In one, the subjects studied a list of 80 words under a levels of processing manipulation with acoustic and semantic processing tasks. After a 24-hour retention interval, they performed a yes/no recognition task; for each item they recognized, they classified their recollective experience as one of remembering, knowing, or feeling. When we analyzed our data the way Tulving and Gardiner did, distinguishing only between remembering and something else, we replicated their essential findings: recognition by remembering was affected by level of processing, but "knowing" as defined by Tulving and Gardiner was not. But when we honored the distinction between remembering and knowing, these experiences were also dissociated by level of processing. In fact, we obtained a double dissociation. Setting aside the category of remembering, which allows the remaining responses to vary freely, there were more recollective experiences of knowing items from the deep processing condition, and more experiences of feeling items from the shallow condition. The pattern for knowing resembled that for remembering, and both were quite different from the pattern for feeling.

In further studies, Kim and I have replicated the dissociation of knowing and feeling by level of processing. We have also shown that repeated exposure increases knowing while it decreases feeling, and that lowering the criterion for recognition increases feeling, but has no effect on knowing. Compared to knowing, feeling is associated with low levels of confidence, and with long response latencies. This is just as it should be, if knowing reflects something like retrieval from semantic memory, and feeling reflects informed guessing.

The point of all of this is that recognition by familiarity is mediated by an experience of perceptual and conceptual fluency which in turn reflects the priming effects of past experience. And it is familiarity -- feeling, not knowing, which seems to underlie successful recognition by amnesics. This is not just speculation on my part. Verfaellie and Treadwell (1993) used Jacoby's (1991) process dissociation framework to study the recognition performance of amnesics, such as revealed by Huppert and Piercy (1976, 1977) and Hirst, Johnson, and their colleagues (Hirst et al., 1986, 1988). Fortunately, I don't have time to explain process dissociation to you, or to defend it against its critics. For our purposes, it is enough to say that process dissociation attempts to separate task performance into its automatic and controlled components -- in this case, recognition by feeling (what Jacoby calls fluency and Mandler calls familiarity) and recognition by remembering (what Jacoby calls recollection and Mandler calls retrieval). For the controls, about 44% of correct recognition judgments were mediated by remembering, and the rest by feeling. Viewed from another perspective, amnesics were impaired only on recognition by remembering. With respect to recognition by feeling, they were indistinguishable from nonamnesics.

More recently, Yonelinas and his colleagues have married the process dissociation procedure to the remember/know paradigm and also to signal detection theory. I am really grateful that I don't have time to explain this menage a trois to you. Anyway, their re-analysis of four previous studies confirms that amnesia has a big impact on recognition by remembering, but a smaller effect on recognition by familiarity. In a new experiment, they found that amnesia impaired both remembering and feeling compared to controls. But the important point was that there was some recognition spared in the amnesics, as we find so often to be the case. And, as would have been expected, almost all of it was based on a feeling of familiarity -- a feeling of familiarity which is a product of priming.

I should note, for the record, that I think that the underlying assumption of the process dissociation framework, that only two processes contribute to recognition performance, is not quite right, and that the framework needs to be expanded to include knowing, if not believing, as well as remembering and feeling. I also don't think that there's anything automatic about recognition by feeling. The experience of fluency may arise automatically, but the subject takes advantage of it deliberately and strategically. Finally, it's not clear that recognition by familiarity, or fluency, is an unconscious process. To be sure, the person has no conscious recollection of the past, but he or she does have the conscious experience of fluency.

I do, however, share Jacoby's view, which is also Mandler's view, that there are no process-pure memory tests: Every memory task, even free recall, has its explicit and implicit components. It's a person who's doing the remembering, after all, not a brain system or a mental module, and in the course of remembering that person will make use of whatever information is available to decide what happened, and reconstruct a mental representation of the past. Some of that information comes in the form of priming-based intuitions about what might have happened in the past. If the subject has better information available, pertaining to the spatiotemporal context in which the event occurred, and his or her role as the agent or experiencer of the event, these intuitions will play a relatively small role in what is remembered. But in the absence of full recollection, people's intimations of the past will loom large. Something will ring a bell, telling the subject what might have happened. Sometimes the intuition will be wrong; sometimes it will be right. In either case, explicit and implicit memory will interact as the subject tries his or her best to reconstruct the past.


Author Note

Paper presented at a symposium, "Mind, Brain, and Behavior: Memory" at the annual meeting of the American Psychological Association, San Francisco, August 1998. Preparation of the paper was supported by Grant MH-35856 from the National Institute of Mental Health. I thank Marilyn Dabady, Jennifer Dorfman, Michael Kim, Lillian Park, Katharine Shobe, and Heidi Wenk for their collaboration.



Brown, A.S., & Mitchell, D.B. (1994). A reevaluation of semantic versus nonsemantic processing in implicit memory. Memory & Cognition, 22, 533-541.

Challis, B.H., & Brodbeck, D.R. (1992). Level of processing affects priming in word fragment completion. Journal of Experimental Psychology: Learning, Memory, & Cognition, 18, 595-607.

Dorfman, J., & Kihlstrom, J.F. (1994, November). Semantic priming in posthypnotic amnesia. Paper presented at the annual meeting of the Psychonomic Society, St. Louis.

Dorfman, J., Kihlstrom, J.F., Cork, R.C., & Misiaszek, J. (1995). Priming and recognition in ECT-induced amnesia. Psychonomic Bulletin & Review, 2, 244-248.

Gardiner, J.M. (1988). Functional aspects of recollective experience. Memory & Cognition, 16, 309-313.

Hirst, W., Johnson, M.K., Phelps, E.A., Risse, G., & Volpe, B.T. (1986). Recognition and recall in amnesics. Journal of Experimental Psychology Learning, Memory, & Cognition, 12, 445-4

Huppert, F.A., & Piercy, M. (1977). Recognition memory in amnesic patients: A defect of acquisition? Neuropsychologia, 15, 643-652.

Jacoby & Dallas. (1981). On the relationship between autobiographical memory and perceptual learning. Journal of Experimental Psychology: General, 110, 306-340.

Kihlstrom, J.F. (1980). Posthypnotic amnesia for recently learned material: Interactions with "episodic" and "semantic" memory. Cognitive Psychology, 12, 227-251.

Kihlstrom, J.F. (1997). Consciousness and me-ness. In J. Cohen & J. Schooler (Eds.), Scientific approaches to the question of consciousness (pp. 451-468). Mahwah, N.J.: Erlbaum.

Kihlstrom, J.F., Shames, V.A., & Dorfman, J. (1996). Intimations of memory and thought. In L. Reder (Ed.), Implicit Memory and Metacognition (pp. 1-23). Mahwah, N.J.: Erlbaum.

Kihlstrom, J.F., Kim, M., & Dabady, M. (1996, November). Remembering, knowing, and feeling in episodic recognition. Paper presented at the annual meeting of the Psychonomic Society, Chicago.

Kim,. M., & Kihlstrom, J.F. (1997, November). Exposure, confidence, and recollective experience in episodic recognition. Paper presented at the annual meeting of the Psychonomic Society, Philadelphia.

Mandler, G. (1980). Recognizing: The judgment of prior occurrence. Psychological Review, 87, 252-271.

Schacter, D.L. (1987). Implicit memory: History and current status. Journal of Experimental Psychology: Learning, Memory, & Cognition, 13, 501-518.

Tulving, E. (1985). Memory and consciousness. Canadian Psychology, 26, 1-12.

Verfaellie, M., & Treadwell, J.R. (1993). Status of recognition memory in amnesia. Neuropsychology, 7, 5-13.

Yonelinas, A.P., Kroll, N.E.A., Dobbins, I., Lazzara, M., & Knight, R.T. (1998). Recollection and familiarity deficits in amnesia: Convergence of remember/know, process dissociation, and ROC data. Neuropsychology, xx, xxx-xxx.


This page last updated 04/08/10 02:58:38 PM.