Copyright © 2007-2017 Russ Dewey
The field of memory research Ebbinghaus founded became known as the verbal learning tradition. The materials could be nonsense syllables, words, or sentences. This label persisted through the 1970s.
Most of the classic memory research techniques were developed during this 90 years period between Ebbinghaus and the end of the verbal learning era. The change was marked by a name change by the primary journal in the field, the Journal of Verbal Learning and Verbal Behavior (JVLVB) to the Journal of Memory and Language in 1984.
"Memory and Language" is a good summary of what the field of verbal learning concentrated upon. Basic memory tests usually involved language until researchers started to test memory for images and other stimuli in the cognitive era. Then the same techniques used for language (recall and recognition testing) were easily applied to other materials such as images.
The simplest memory technique is recall testing. To recall something is to retrieve it from memory upon request.
Several different types of recall tests were developed during the heyday of verbal learning theory, from about 1915 to 1965. These included serial learning, paired-associates learning, free recall, and cued recall.
The techniques are still used today, although not usually with nonsense syllables. We will briefly review each on this page.
Serial learning occurs when you learn something in a fixed sequence or rigid serial order. A child learning the alphabet must remember the letters in exact order.
A person learning a route through a complex environment must remember a series of turns in order. Many procedures such as baking a cake or tying a shoelace involve actions performed in a certain order.
What is serial learning? What are examples of situations that require serial learning?
In a serial learning study from the nonsense syllable era, you might have been asked to memorize this list:
Later, you would be asked to recall the items in the same exact serial order. That makes it a test of serial learning.
How did early psychologists explain serial learning?
In the old days of verbal learning research, from the 1880s to the 1950s, most psychologists thought serial learning was a process of forming little links (associative bonds) between items in a series. After the first item in the list, each subsequent item served first as a response, then as a stimulus for the next item. Psychologists diagrammed the process like this:
John B. Watson, the famous behaviorist, believed this was how rats learned to run a maze. He thought the rat connected stimuli such as the sight of a particular wall with a response such as making a right turn.
The way Watson conceived it, each response exposed the rat to a new stimulus, provoking a new response. When the whole chain of S's and R's was linked, the rat could run the maze without errors.
However, serial learning does not really work like this, either for rats or humans. Rats (as we now know, over a century later) use cognitive maps that include location information, when running a maze.
Human subjects, reflecting their mastery of language, often use very complex information processing to remember even a simple list of nonsense syllables. Miller, Galanter and Pribram (1960) pointed this out:
If you ask a man who has just memorized his first list of nonsense syllables to tell you what he did in order to master the list, he will have quite a lot to say.
...Now that first nonsense syllable, BOF, was just plain remembered the way it came, but the second one reminded him of "XAJerate," the third one turned into "MIBery," and the fourth turned from ZYQ to "not sick."
So he had a kind of sentence, "BOF exaggerates his misery because he is not sick," instead of the cryptic BOF, XAJ, MIB, ZYQ, and he could imagine a hypochondriac named BOF who continually complained about his health. (Miller, Galanter, and Pribram, 1960, p.126)
Ebbinghaus would presumably not approve of this. Ebbinghaus was trying to eliminate the influences of meaning, by using nonsense syllables. However, as Miller, Galanter, and Pribram pointed out, most humans naturally use such associations to support memory.
Note the date on the Miller, Galanter and Pribram example: 1960. That is about the time computers began to appear.
Computers made it obvious that any information going into a system had to be encoded or transformed in some way so the system could take it in. Miller, Galanter and Pribram were among the earliest memory researchers in the late 1950s who realized this insight would apply to human memory.
The word encoding refers to interpretation of a stimulus. It happens in all forms of memory. Nothing gets into an information processing system unless it is encoded in some way, first.
This insight could not be avoided, in the early days of computing, because to use the new giant electronic brains (a 1950s term for computers) researchers had to punch all their data in to IBM cards. Those were punch cards used by IBM, the largest computer manufacturer, a company that previously made business machines such as cash registers.
Computers appeared in the most technologically advanced university research centers in the mid- to late-
Computers highlighted the concept of information processing. Thoughtful experimental psychologist soon realized that humans, too, were information processors.
When it became clear that nothing gets into an information processing system without being encoded in some way, researchers started realizing the same thing must be true of humans. To remember something, a person first had to take it in, encode it, and that required some sort of interpretation.
Often the errors a subject makes were an important clue to how the subject encoded the stimulus. Miller, Galanter, and Pribram (1960) continued the story of their hypothetical subject (described above encoding "MIB" as "false misery") this way:
That MIBery-misery association wasn't too good, however, because for two or three trials through the list he remembered MIS instead of MIB. But he finally worked it out by thinking of "mibery" as a new word meaning "false misery."
The example shows how encoding or interpretation of a stimulus item (such as MIB) can be revealed by exploring errors a subject makes. Of course, the experimenter would have to ask the subject why he or she remembered MIS instead of MIB, to find this out.
Therefore investigation of encoding effects depends partly upon introspection: asking participants in an experiment what was going on in their minds. These are not difficult introspections; most people can tell you how they interpreted something while trying to memorize it.
When subjects use elaboration to help remember items on a memory test, researchers call this elaborative encoding. Elaborative encoding helps memory by providing information that aids retrieval.
A second verbal learning technique from the turn of the century was called paired-associates learning. Mary Whiton Calkins, who was president of the American Psychological Association in 1905, invented this technique.
The paired associates method requires a subject to learn pairs of items by forming associations between them. This reflecting the "S-R " assumptions that dominated experimental psychology from about 1900-1950.
According to S-R theory, all learning was portrayed as associations between stimuli and responses. In paired-associates learning, the first item is the stimulus, the second item is the response.
In paired-associates learning, a subject studied lists of word pairs. The list was considered memorized when the subject could respond to any stimulus from the list with its associated response.
That was supposed to resemble real-life situations. For example, a person might respond to a person's face with memory for that person's name.
What is the paired-associates method?
In a paired-associates experiment, you might see a list like this:
After you inspect the list and attempt to memorize it, the experimenter presents a stimulus item (like 7) and you try to supply the response item (CSL). Then the experimenter might present the stimulus 4 and you would answer VNR.
You must learn the association between each stimulus and response. The pairs could be presented in any order.
Suppose you were trying to defeat Ebbinghaus's original intentions again by using meaning or elaborative encoding to memorize nonsense syllables. The first step might be to think of some word resembling the nonsense syllable.
For example, CSL might remind you of castle. If the pair was 7-CSL and you knew of the classic story, "The House of Seven Gables," you might imagine an ornate house with seven gables (seven points on the roof). This would be another case of elaborative encoding.
What are "natural language associations"?
Researchers use the term natural-language associations to refer to examples like this and the previous one (involving "mibery"). Both involve words and sentences from ordinary language.
Different people make different associations. Research suggests that people do best with their own associations, rather than those supplied by an experimenter.
Free recall tests are, in some ways, the simplest form of memory test. A person inspects a list of items then (after a retention interval) tries to recall the items in any order. The items can be letters, words, trigrams, sentences, or longer passages such as stories.
What is free recall?
In free recall studies, subjects are free to use any strategies they want. However, memory researchers found that subjects tended to organize the items they were memorizing. For example, suppose a subject had to memorize this list:
After a retention interval of 15 minutes, the subject might recall the items as follows.
"Nixon, Bush, Carter, Red, Green, Blue"
Because this is a free recall experiment, the subject is allowed to recall the items in any order.
As early as the 1940s, psychologists discovered that people tended to recall clusters of items from the same category, during free recall. This was called clustering, and it seemed to occur automatically. People were not necessarily aware they were doing it.
In this case, the subject created one cluster of president names, one cluster of color names. The subject probably encoded BUSH as a president's name rather than a shrub. The encoding was biased by the presence of other president's names.
What is clustering? Does it require conscious intention?
Errors in free recall can also be revealing. Notice, in the above example, the substitution of blue for orange.
How can errors be revealing?
This is a typical error in recall. Again, the error reveals how the subject encodes or interprets items in the list. In this case the error confirms that the subject was using color names as an organizing theme.
Cued recall is like free recall except subjects are given hints (cues) at the time of recall. The cues are supposed to help the subject recall the memorized items.
However, to be effective, the cue must relate to the way a subject interpreted the item in the list. If you interpret plane as flat surface, then a cue like flying will not help you remember it later. If you interpret plane as airplane, then the cue flying will probably help you recall it.
How do cued recall experiments work?
Cued recall is useful for detecting unconscious or automatic aspects of encoding. For example, using cued recall, researchers discovered a phenomenon called spontaneous trait inference.
Participants could be given a sentence like "The librarian carried the old woman's groceries across the street." The word helpful then aided recall (Winter & Uleman, 1984).
The word helpful did not appear in the original sentence. Subjects inferred that the librarian was helpful, so this became part of their encoding: part of their memory for the sentence. Therefore "helpful" was an effective prompt to recover the memory.
How can cued recall be used to demonstrate spontaneous trait inference?
This was called spontaneous trait inference because subjects did it spontaneously (automatically, without instruction). It involved an inference (a logical assumption that a person aiding an older lady is helpful). It also involved a trait of a person (helpfulness).
Spontaneous trait inference is one of many complex cognitive processes illuminated by cued recall experiments. Researchers showed that people automatically elaborated on the meaning of a simple sentence, making logical inferences that influenced memory recall later.
Miller, G. A., Galanter, E., & Pribram, K. H. (1960) Plans and the Structure of Behavior. New York: Henry Holt and Company, 1960.
Winter, L. & Uleman, J. S. (1984). When are social judgments made? Evidence for the spontaneousness of trait inferences. Journal of Personality and Social Psychology, 47, 237-252.
Write to Dr. Dewey at firstname.lastname@example.org.
Copyright © 2007-2017 Russ Dewey