Underwood, Benton (1915-1994)

views updated


Benton J. Underwood was one of the preeminent leaders in the postwar development of research on the acquisition and retention of verbal materials (Keppel, 1997, p. 469). He studied verbal learning and memory in the 1940s at the University of Missouri, then later at the University of Iowa, under such important figures as Arthur W. Melton (Missouri), John A. McGeoch, and Kenneth W. Spence (both at Iowa). In 1946, Underwood took a teaching position at Northwestern University in Evanston, Illinois, where he remained until his retirement in 1983. Over four decades he did groundbreaking work in associative learning, verbal discrimination, transfer of training, distribution of practice, interference and forgetting, and the composition of memory.


Hermann Ebbinghaus carried out the first systematic study of verbal learning and memory in 1885. Acting as both experimenter and subject, he learned lists of nonsense syllables (e.g., cak, roq) and then tested his retention of them. His efforts provide our first picture of an empirically generated forgetting curve. Ebbinghaus's curve revealed substantial forgetting in the hours immediately following original learning. Why was there so much forgetting so soon after learning? This puzzle Underwood later helped to solve.

At the start of the twentieth century, William James's work was the chief influence on American psychology. Unlike German psychologists, who emphasized the mind's structure and had given birth to a scientific psychology, James stressed the mind's activity. His ideas, combined with those of British philosophers and the American spirit of doing and acting, created a movement called American Functionalism. A student of Underwood's once commented: "To my mind, the term functionalist and the name Underwood are synonymous" (Freund, 1998, p. 318).

The functionalist perspective within the newly developing field of psychology approached problems of learning and memory in associationistic terms. How are associations between verbal items learned and remembered? As Underwood commented, "I am an incurable associationist" (1982, p. 8). It seemed obvious to him that in innumerable instances associations played a dominant role in our learning and retention of verbal material: When did Christopher Columbus discover the new world? Who is the current vice president? Clearly, to answer such questions we must learn an association between initially unrelated facts (e.g., between a name and date). The experimental task given to subjects typically involved paired-associate learning: the presentation of pairs of items to a subject charged with learning an association between the members of each pair to facilitate the recall of the second item upon presentation of the first.

Another important characteristic of the functionalist perspective was its use of stimulus-response (S-R) language borrowed from behaviorism, which dominated American psychology until the 1950s. Paired-associate learning, for example, was described as learning an association between a stimulus (the first member of the pair) and a response (the second member). Principles of classical conditioning were incorporated into explanations of human learning and retention of verbal material. Forgetting of verbal material, for example, was discussed in terms of extinction and spontaneous recovery of associations, a language perfectly in tune with that of the behaviorists. Other salient characteristics of the functionalist orientation to psychology include an insistence on modest theory building, a close interplay between theory and data, and an interest in analyzing phenomena in terms of their component processes or parts.

Learning and Memory

In 1932 McGeoch argued that forgetting was due to the intervention of events that occured after the acquisition of information and was not simply a matter of decay or disuse. Forgetting caused by a learner's activities in the interval between original learning and a test of retention is called retroactive inhibition. Its primary mechanism was thought to be response competition. On this view, forgetting occurs when new learning competes with the original learning. In 1959, Underwood provided evidence for an additional mechanism, that of unlearning. He showed that original learning actually became unavailable for recall as new learning proceeded.

Nevertheless, additional research suggested that the more powerful factor producing forgetting was proactive inhibition. Proactive inhibition occurs when information learned before the target material interferes with retention (e.g., an old telephone number intrudes on your attempt to recall a more recent number). Because memory researchers often had their subjects learn and remember many lists, there was an opportunity to examine forgetting as a function of the number of prior lists acquired by subjects in the learning laboratory. In 1957, in a classic piece of detective work, Underwood showed that retention of a single list of items after twenty-four hours decreased as a function of the number of prior lists that had been learned (see Figure 1). Retention of a single list without prior learning was about 75 percent, not just 25 percent, which had been a conclusion drawn from Ebbinghaus's forgetting curve. Ebbinghaus was affected by the proactive inhibition resulting from learning and recalling many, many lists. In later years Underwood and his students demonstrated how proactive interference operates even in very short-term memory situations and provide further accounts of possible sources of proactive interference in long-term retention of verbal material.

An associationistic perspective led Underwood to conclude that associations of common words will likely occur implicitly during a learning task. An adult learning a list of words such as sugar, bitter, and candy will likely implicitly think of the word sweet. In 1965, Underwood showed that these implicit associative responses, or IARs, played a role in producing false recognition. For example, learners who experience words related to sweet, such as the above, later come to falsely state that they had actually experienced the word sweet. Underwood demonstrated a role for IARs in many different learning and memory tasks.

Attributes of Memory

Consider a simple learning task. A list of word pairs is presented. Your task is simply to remember which word in each pair is the right one (for example, as designated by the experimenter by underlining). When the word pairs are shown again (with no identification of right items) can you remember which is the right one? Of course you can, at least after a few learning trials. But how do you do it? In 1966, Underwood and his students presented a frequency theory of verbal discrimination learning. It is a beautiful example of the functionalist approach. The theory is modest in scope (accounting mainly for verbal discrimination decisions), closely tied to the data (making clear, easily tested predictions), and permits analysis of a larger phenomenon (verbal discrimination) into its constituent processes. The theory posits that subjects make decisions in verbal discrimination tasks based on a subjective frequency differential between right and wrong members. This situational frequency differential is built up through representational responses (a perceptual processing of the item), rehearsal responses, and implicit associative responses.

This theory has survived many tests and has been extended to other situations, accounting, for example, for our ability to perform on a recognition memory test. Consider what sometimes happens on a multiple-choice exam. Initially, one alternative looks right (familiar), but as we consider the other alternatives, our confidence decreases. Frequency theory suggests that the initial verbal discrimination is based on a subjective difference in frequency that arises from our prior study of the right item; but that differential is lost as we raise the frequency (familiarity) of other alternatives.

Underwood's thinking about frequency as an attribute of memory led to a consideration of the composition of memory. Or, as Underwood asked in 1969, "Of what does a memory consist?" (p. 559). His answer was that memory, our record of an event, is a collection of attributes. There is no corpus or body of memory that can be recalled directly. Memory attributes both aid discrimination (e.g., frequency) and retrieval (e.g., IARs). Many everyday experiences provide evidence of such a conceptualization. Tip-of-the tongue experiences, for example, sometimes involve an ability to remember letters of a word (orthographic attribute) when we can't remember the whole word. Many students have had the experience of being able to remember where on a textbook page an answer was studied but not being able to remember the item itself (spatial attribute). Underwood's attribute theory helps to explain learners' performance in many tasks and, along with other multidimensional theories, has enhanced our understanding of memory.

Underwood published more than 200 scientific articles, books, chapters, and monographs, and his work elicited a plethora of awards and honors, including election, in 1970, to the National Academy of Sciences. He was a distinguished teacher and superb methodologist whose research and textbooks inspired the work of countless students and colleagues.



Barnes, J. M., and Underwood, B. J. (1959). "Fate" of first-list associations in transfer theory. Journal of Experimental Psychology 58, 97-105.

Duncan, C. P., Sechrest, L., and Melton, A. W., eds. (1972). Human memory: Festschrift in honor of Benton J. Underwood. New York: Appleton-Century-Crofts.

Ekstrand, B. R., Wallace, W. P., and Underwood, B. J. (1966). A frequency theory of verbal discrimination learning. Psychological Review 73, 566-578.

Freund, J. S. (1998). Benton J. Underwood: A tribute of memories. In G. A. Kimble and M. Wertheimer, eds., Portraits of pioneers in psychology, Vol. 3. Mahweh, NJ: Erlbaum.

Keppel, G. (1997). Benton J. Underwood (1915-1994). American Psychologist 52, 469-470.

McGeoch, J. A. (1932). Forgetting and the law of disuse. Psychological Review 39, 352-370.

Postman, L. (1972). The experimental analysis of verbal learning and memory: Evolution and innovation. In C. P. Duncan, L. Sechrest, and A. W. Melton, eds., Human memory: Festschrift in honor of Benton J. Underwood. New York: Appleton-Century-Crofts.

Underwood, B. J. (1949). Experimental psychology. New York: Appleton-Century-Crofts.

—— (1957). Interference and forgetting. Psychological Review 64.

—— (1957). Psychological research. New York: Appleton-Century- Crofts.

—— (1965). False recognition produced by implicit verbal responses. Journal of Experimental Psychology 70, 122-129.

—— (1969). Attributes of memory. Psychological Review 76, 559-573.

—— (1982). Studies in learning and memory: Selected papers. New York: Praeger.

Zechmeister, E. B., and Nyberg, S. E. (1982). Human memory: An introduction to research and theory. Belmont, CA: Wadsworth.

Eugene B.Zechmeister