Psi man mascot

Secondary Memory

Atkinson and Shiffrin depicted long-term memory as one box in the mid-1960s. After that, psychologists found several distinct varieties of secondary or long-term memory:

The first two, episodic memory and procedural memory, use distinct biological systems within the brain. Episodic memory involves the hippocampus and adjacent areas of the temporal lobe; procedural memory involves the cerebellum, the "small brain" at the back of the skull.

Implicit memory can occur in any brain circuit; it is memory that takes place without executive control. Explicit memory requires some mental effort or coordination by executive processes.

Episodic Memory

Memory researcher Endel Tulving (1972) coined the term episodic memory. This memory of single episodes of your life can also be called autobiographical memory or event memory.

Episodic memory binds together all the different features of an experience (sights, sounds, emotions, thoughts) into a single memory. The hippocampal circuit is presumed to be specialized for this. It draws from all the other regions of the brain, knitting together a unified representation we experience as memory for one event.

What is the binding problem?

Episodic memory is like a road from the past to the future, consisting of a chain of episodes or experiences. Every step along the path has an autobiographical tag on it: a label that says, "This happen­ed to me; this was an event in my life."

Episodic memories are constructions, so they can have errors. They generally feel real, accurate, and true even if they are confabulations (fictional constructions mistaken for accurate memories).

The only way we know if a retrospective self-report is confabulated is to find objective evidence relating to it. Most of the time we do not have ready access to evidence about past events, although when we look through old family albums or discover school work from childhood, we may see evidence of events we do not remember.

Most of the time we simply accept that what we remember about our lives is accurate, and that works out fine. However, psychologists have no trouble demonstrating that people can have false memories, and as we saw in the discussion of leading questions and hypnosis, false event memories are easily induced.

Episodic memory is clearly a distinct system because it can be lost after brain damage (see next page). It can also be temporarily disabled in the syndrome known as dissociative amnesia.

Dissociative amnesia is classic amnesia as seen in old-time movies. People who experience a trauma may react, in rare cases, by forgetting who they are.

A person with dissociative amnesia forgets personal information: memory linked to their identity. In other words, they lose episodic memory.

However, they retain memory for factual knowledge. Such a person could tell you that 2+2 equals 4, or what politician is leading the country.

What happens in cases of dissociative amnesia?

Declarative and Procedural Memory

Declarative knowledge like "2 + 2 = 4" or "Hong Kong is part of China" is built up through repeated encounters with information. It is not linked to a single event. This distinguishes such knowledge from episodic memory.

Tulving's term for non-episodic memory was semantic memory and you will still see that term used in memory literature. The label has gone out of favor with some psychologists because "semantic" means "related to word meanings" and this type of memory involves more than words: it involves facts and knowledge in general. The term declarative memory can be used instead, or factual memory, or world knowledge.

Why has the name "semantic memory" gone out of favor with some psychologists?

Procedural memory is specifically memory for sequences of events, processes, and routines. Remembering how to tie your shoes, make a layup on a basketball court, or make tea requires procedural memory.

The fact that procedural memory was distinct from other types dawned on psychologists in the mid-20th Century, when they observed that patients receiving ECS (electroconvulsive shock) for severe depression would show forgetting of factual information presented to them in the preceding days. They showed no loss of procedural skills they practiced during that time.

Similar evidence came from the famous patient Henry Molaison or "H.M.," who completely lost his ability to form new long-term memories after surgery. We will discuss his story later.

What is evidence procedural knowledge is distinct from declarative memory?

Although H.M. could form no new event memories, he could still learn motor skills such as mirror tracing (tracing within the outlines of a star shape, seen in a mirror). Mirror tracing requires procedural memory: memory of how rather than memory of what.

H.M. improved in mirror tracing even though he never remembered the practice sessions. This impressed upon psychologists of the mid-1960s that procedural memory must be distinct from other long-term memory processes.

Implicit vs. Explicit Memory

Certain forms of memory do not require conscious executive control. An example is remember­ing how to brush your teeth. You probably do not think about it.

Automatic forms of memory are called implicit memory. They are usable even if a person suffers severe brain damage from organic brain syndromes such as Alzheimer's Syndrome.

What is the difference between implicit and explicit memory?

People who are drunk, or who suffer from organic brain syndromes, may perform very poorly on memory tasks requiring conscious control. Those are called explicit memory tasks. An example of an explicit memory task would be reciting everything you remember reading on the page about Ebbinghaus.

By contrast, implicit memory tasks are automatic responses. They are drawn out of a person by circum­stances, elicited by a situation, like clapping after a per­formance.

In the following columns, tasks on the left are examples of implicit memory that do not require executive control. Demented or intoxicated humans can still perform them.

Tasks on the right are examples of explicit memory that do require some conscious control. They are performed poorly by people who are severely brain-damaged or drunk.

Tasks requiring implicit memoryTasks requiring explicit memory
Turning on a TVRemembering what you watched on TV
Reading reversed textPaired associates learning
Doing a word completion taskIdentifying heads of state
Singing part of a familiar songWriting a term paper

Items in the left column are all indirect forms of memory. They do not require a conscious strategy for retrieving information.

If you once learned to read text that is printed backwards, chances are you will be able to do it later, even if you suffer a brain disorder. The same is true of the other tasks on the left.

How does implicit memory help old people who stay in their own homes?

The importance and power of implicit memory helps explain why old people are often more comfort­able and capable when they stay in a familiar place. Years of living in the same rooms produces implicit memory such as knowing where to find a broom and other useful knowledge and skills.

When an old person is put in a nursing home or other unfamiliar environment, the same person may seem disabled, because none of the old automatic routines work in the new environment.

What was discovered in the study of "alcohol amnesia?

Hashtroudi, Parker, DeLisi, Wyatt, and Mutter (1984) studied the effects of alcohol on implicit and explicit memory. Ninety-six male volunteers between the ages of 21 and 35 were recruited for the all-day experiment (they were kept at the lab until their blood alcohol returned to zero).

The researchers found that alcohol intoxication had effects similar to brain injury. It damaged explicit memory but not implicit memory.

The subjects were divided into 6 groups of 16. Three groups received alcoholic drinks; three groups received placebos. Each received 4 drinks in 40 minutes, then the memory tests began.

Subjects saw a list of words for 2 seconds each. They performed an arithmetic task for 5 minutes, then they were tested on one of three memory tests: (1) recalling the words, (2) recognizing the words, (3) identifying the words in degraded form.

words are blurry at first then progressively clearer
Degraded words progressively filled in

The figure shows some degraded words used in the experiment. The subject had to say the word as quickly as possible.

First they saw the word in severely degraded form (top. Then more and more of the word revealed (as shown top to bottom).

Alcohol-intoxicated subjects performed just as well as sober subjects on the last two tasks: recognizing the degraded words and recognizing other words in a Yes/No recognition test.

Both forms of recognition are examples of implicit memory, because a person does not have to perform any elaborate or deliberate mental activity to recognize a word. You see the word, later on you recognize it, and it is all somewhat automatic and effortless.

What tasks were affected by alcohol, and what were not?

However, on the recall task, a big difference appeared. Drunk subjects recalled only half as many words as sober subjects.

This was a task that involved explicit memory. Subjects knew they would be asked to recall the words later, so they probably tried to memorize them during the initial presentation of the list. Sober subjects were better at this.

Implicit memory occurs under the surface all the time. As you explore the environment, you absorb information about it without trying.

If somebody asks you where a certain missing object is located, you may have that knowledge even though you did not make a conscious effort to memor­ize the object's location. No effort is required to retrieve this information if you have it.

Because implicit memory operates without conscious intervention, preschool children are sometimes as good as adults at tasks requiring only implicit memory. Ask a 3 year old where a favorite toy is, and the child is as likely as you are to remember.

Ask a 3 year old to hold a series of numbers in memory and he or she will not perform as well. Even if a child can count to 10, he or she will not know how to rehearse a list of numbers to keep them in primary memory. That explicit memory strategy is learned in later years.

With what type of memory are preschool children just as good as adults?

Implicit memory is very pervasive or widespread in human cognitive processing. Robert G. Crowder of Yale University pointed out, "Most researchers now agree that implicit memory is more influential than explicit, conscious memory" (Bower, 1990).

Priming in memory research

A technique called priming can demon­strate implicit memory. A person who sees the word yellow will be slightly faster to recognize the word banana as a word. This happens because the words yellow and banana are closely associated in memory.

a network of related words
A semantic network

Researchers envision a network of word meanings or semantic network some­what like the diagram. The distance between words indicates the frequency with which the words are associated in everyday life.

What is "priming"? What is a "semantic network"?

Because of these associations, acti­vating one node of the network (showing the person one word) warms up or primes nearby words, speeding retrieval. This effect lasts about 30 minutes after exposure to the priming word.

Different tasks produce different durations of priming effects. Tulving, Schacter, and Stark (1982) used a word completion task. For example, _YS__RY is the word mystery with several letters removed. Participants were given the fragmentary version and asked to figure out the word.

Seven days later, the participants were tested again, and some of the same words were used. They were faster to figure out words they had seen 7 days earlier, even if they could not recognize the words as having occurred in the first phase of the experiment.

No conscious strategy is required to show priming effects. Brain-damaged and intoxicated people show the same priming effects as other people.

Are priming effects implicit or explicit?

The example used above to illustrate implicit vs. explicit memory, using degraded words, was also a form of priming. If participants had seen the word for a two seconds, five minutes earlier, they recognized it faster when presented in degraded form.

In that case, the experimenters were interested in seeing whether the priming effect would occur equally in drunk and sober subjects. It did.

Priming helps normal reading. Words seen ahead of the fixation point of the eye (in peripheral vision) are activated in semantic memory (warmed up) so when the eye fixates upon them, their meanings are available faster.

How does priming normally help language comprehension?

Similarly, in conversation, if you hear somebody say, "I ate a yellow" [followed by a muffled word that sounds like "nee-nah"] you might well hear "I ate a yellow banana" because you have a semantic network like the one in the diagram.

The word banana is activated by its association to the word yellow, so you easily retrieve it even if the stimulus is partial or degraded. The memory retrieval is automatic, evoked by the situation, so this is an example of implicit memory.

Memory as a Construction

In the time of Ebbinghaus in the 1880s until the encoding revolution of the 1960s, memories were considered to be like snapshots that could be retrieved whole.

The physical basis of a memory was called an engram. Engrams were assumed to be stored in some particular place in the brain. That sounds a bit like event memory, which is removed if the hippocampus is removed. As we have seen, there are other types of memory.

With the advent of the encoding per­spective, scientists began to consider storage requirements for human memory. They realized it would be far more efficient to store the instructions for cognitive events, rather than whole productions.

Why are "instructions" stored instead of whole memories?

To understand the distinction, think of how difficult it would be to store thousands of cakes. It would require a small warehouse, and even if the cakes were refrigerated, they would deteriorate over time.

But thousands of recipes (instructions for making cakes) can be stored in a filing card box or small computer database. In this compact form the information can endure for a long time unchanged.

When it is time to "remember" a cake, the recipe is retrieved and the cake is baked again, following the instructions. This is much more efficient than storing whole cakes, and it results in a fresh product each time. The same is true of memories.

Memories are not stored whole. Instead, we store data used to generate an image or word or scene or story from our past. Each time we remember something, we produce a fresh construction.

Like a chef baking a cake, we can make errors in reconstructing memories. Sometimes we introduce changes that result in a totally different product.

Memory errors are commonplace. We have no easy way to tell if memories are accurate, except double-checking objective facts. A memory that feels like a vivid personal experience may, in fact, be based on second-hand information.

A student gives this example:

Many years ago, my family and I lived in Mexico City, Mexico. We used to spend Christmas and other holidays in Acapulco, the site where the famous, daring cliff divers can be seen.

A few years ago, I was relating a very real experience to some of my friends, concerning the death of one of the divers and my witnessing it. The diver had gracefully flung himself off the cliff, misjudging the jagged rocks lining the water, and died on impact.

My parents overheard the conver­sation and later told me that although the event occurred, I had not been present at the time.

Yet it all seemed so vivid to me: sitting on my Father's shoulders, peering over a crowd of tourists, massive cliffs, crashing waves, that diver, and the endless blue Pacific Ocean.

I'm sure this is an example of a construction that feels like a memory. Whether I remember this through a newspaper or the TV I don't know, but to this day it all seems so real. [Author's files]

How does the student's essay illustrate a commonplace occurrence?

Memory errors are a vexing problem for therapists who deal with recovered memories. People sometimes remember traumatic incidents from earlier in life, but sometimes these memories are erroneous. It is very difficult to tell the difference unless objective evidence is available.

Memory problems are also problematic for lawyers. The advent of DNA testing revealed many cases in which eyewitness identification resulted in wrongful imprisonment.

This problem is not easy to solve. Even when people are asked to take special care to avoid memory distortions, they still make them (Winkelspecht and Mowrer, 1999).


Hashtroudi, S., Parker, E. S., DeLisi, L. E., Wyatt., R. J., & Mutter, S. A. (1984) Intact retention in acute alcohol amnesia. Journal of Experimental Psychology: Learning, Memory, and Cognition, 10, 156-163.

Tulving, E., Schacter, D. L., & Stark, H. A. (1982) Priming effects in word-fragment completion are independent of recognition memory. Journal of Experimental Psychology: Learning, Memory, and Cognition, 8, 336-342.

Winkelspecht, S. M. & Mowrer, R. R. (1999) Memory distortion: Can accurate memory be preserved? Psychological Record, 49, 137-144.

Write to Dr. Dewey at

Don't see what you need? Psych Web has over 1,000 pages, so it may be elsewhere on the site. Do a site-specific Google search using the box below.