1906—1978 · arthur weever melton august 13, 1906-november 5, 1978 by michael i. posner arthur...

n a t i o n a l a c a d e m y o f s c i e n c e s

Any opinions expressed in this memoir are those of the author(s)and do not necessarily reflect the views of the

National Academy of Sciences.

a r t h u r W e e v e r m e l t o n

1906—1978

A Biographical Memoir by

m i c h a e l i . p o s n e r

Biographical Memoir

Copyright 1992national aCademy of sCienCes

washington d.C.

ARTHUR WEEVER MELTON

August 13, 1906-November 5, 1978

BY MICHAEL I. POSNER

ARTHUR MELTON was already a full professor of psychol-ogy, a brigadier general in the Air Force Reserve, and

the editor of the Journal of Experimental Psychology when Ifirst met him as a new graduate student at the University ofMichigan in 1959. It was only the second year of his re-turn to academia after having served as technical directorof the Air Force effort in human personnel for eight years.Of course, as is typical of graduate students, I assumed hehad always been at Michigan, although I remember nowreferences to his time at Ohio State. I mention this be-cause I was neither a contemporary nor, still less, in a posi-tion equal to Melton's, and thus I could not know him asthose who were his close friends did. Rather, I knew himas an imposing force in the field, as a teacher of consider-able influence, and as a role model even for those whowere not direct students of his.

More personal memoirs of Melton have been written byBenton Underwood, who was Melton's student at Missouri;by Robert Daniel, who also knew Melton at Missouri; byFrank Geldard, who knew Melton during the period of hisAir Force work; and by Jim Greeno and Wilbert McKeachie,who were colleagues and coworkers at Michigan. Thesewriters divide Melton's work into his early academic period

315

316 BIOGRAPHICAL MEMOIRS

at Missouri, his Air Force period, and the later period inwhich the Human Performance Center was built at Michi-gan. Perhaps because my daily involvement with Meltonwas much less extensive than theirs or because fourteenyears have passed since his death, it seems possible to viewthese separate periods more as a whole and see the conti-nuity of his developing ideas about psychology in relationto changes in the field.

Recent histories of experimental psychology have oftenidentified the study of information processing entirely withdevelopments stemming from the efforts to use computersto stimulate higher-level human functions. Indeed, the workof Herbert Simon and Allen Newell, among others, hasbeen very influential in developing cognitive science. TheMichigan Human Performance Center, together with theApplied Psychology Unit in Cambridge, Massachusetts, rep-resented a different tradition that also contributed to theview of the human as an information processor. The Michiganapproach arose from the functionalist tradition. It empha-sized quantitative empirical relationships. These includedFitts' law relating movement speed to its distance and ac-curacy of termination, the Hick-Hyman law relating reac-tion time to the number of alternatives, and Melton's ownempirical relation between the recallability of an item andthe spacing between its repetitions:

Analogies between the human mind and technology changeover the years. We are currently undergoing a shift fromvery strong identification of mental processes with serialcomputer systems to an emphasis on parallel activation. Itis quite interesting to see that these new theoretical pro-posals are being tested against the quantitative empiricalgeneralizations that were developed in the functional tra-dition that Melton practiced. The fourteen years that haveelapsed between Melton's passing and this memoir make it

ARTHUR WEEVER MELTON 317

possible to apply to his influence the kind of empiricalcheck that was so characteristic of his approach. The 1987social science citation index listed ten of Melton's publica-tions. His effort to summarize the influence of the secondpresentation of an item on its memory was the most fre-quent, with thirteen citations nearly twenty years after itspublication. I hope this memoir will provide some back-ground for appreciating the nature of Melton's contribu-tion.

EDUCATION

Arthur Melton came to Washington University in St. Louis,Missouri, in 1924 as an eighteen-year-old undergraduatefrom Arkansas. He found there then, as he would now, astrong emphasis on the functional analysis of human be-havior. Functionalism held that a detailed analysis of thetasks of daily life could help us understand how humansrelate to their environment and how the environment mightbe improved to accommodate human nature. John A.McGeoch, with whom Melton worked as an undergradu-ate, performed studies exploring how the distribution ofpractice, rest, and interpolated learning affects the forma-tion and loss of associations. This work led to McGeoch'stextbook Psychology of Human Learning. It was largely a com-pendium of experimental findings of functional relation-ships between variables that could be manipulated by theexperimenter and the changes in performance that theycaused. There was, however, the start of a theory of forget-ting embedded in this material. Information was not lostmerely through a period of disuse, but rather as a functionof the amount and type of learning between the originallearning and testing. Ten years later Melton would make asubstantial experimental contribution to the "interferencetheory" of forgetting.


Melton received his B.A. degree in 1928 and went on tograduate study at Yale University. Four years later he tookhis Ph.D. degree with Edward S. Robinson. Robinson hadcome to Yale from Chicago and was himself a contributorto the functional tradition of verbal learnng.

At the height of the depression, even a psychology de-gree from Yale University was no guarantee of a scarceacademic position. However, Robinson had been involvedsince 1925 in a project supported by the Carnegie Corpo-ration and the American Association of Museums, and Meltonwas employed for three years in studies leading to impor-tant monographs on the actual behavior of visitors to mu-seums and a comparison of methods of instructing childrenin science museums.

Two things of great importance for Melton's future workemerged from that assignment. First, he learned that psy-chology could make contributions to real-life problems.Second, he began to explore what process might producethe functional relationships he observed. Consider as anexample Melton's measurement of the time that visitorsremained in front of the displayed paintings (dwell time).He plotted the dwell time for each month of museum op-eration. The dwell time was at a minimum in the summermonths and reached a peak in the winter. This was trueboth for weekday and Sunday visitors. The reason for this,Melton speculated, was not in the behavior of individuals,as would be the case if, during the summer season, eachperson reduced the time he or she stayed indoors, butrather in the type of visitor attracted to the museum in thewinter and summer seasons. While Melton was frustratedin not being able to prove this speculation in the museumsetting, he was able to show that the visits of relativelyuninterested visitors increased toward summer, presum-ably because of the ease of getting there, while those of


relatively interested visitors declined, possibly because ofcompetition from other activities. These studies illustratedboth the functional approach of looking at consistent gen-eralizations from systematic empirical data and also an ef-fort to go beyond the functional relationship toward thedevelopment of a process that might underlie the functionalrelationship. The scarcity of such process explanations wasone reason why the functional approach often seemed souninteresting. In addition, these studies illustrated thatsystematic observation of natural behavior could be appliedto the actual implementation of museum exhibits.

PRE-WAR RESEARCH

In 1932 Melton's former mentor John McGeoch left thechairmanship of the Psychology Department at the Universityof Missouri, Columbia. With the strong recommendationof Robinson, Melton assumed the position as head of thePsychology Department at the University of Missouri at theage of 29.

It was the development of the M.A. program at Missourithat ensured Melton's reputation as an administrator. Meltonand the four members of the Psychology Department atthe University of Missouri produced a small and thrivingintellectual home for a number of outstanding psycholo-gists, including Alvin Liberman (a member of the NationalAcademy of Sciences), Benton Underwood (a member ofthe National Academy of Sciences), Robert Malmo, DavidC. McClelland, Richard Q. Bell, and Moncrieff Smith,

During this time, Melton also showed how a functionalapproach to human learning might be combined with pro-cess ideas to reveal the operations that take place in themind of the subject but that are not directly observable.The importance of Melton's paper on the role of interpo-lated learning can only be understood in light of the ex-


treme inhibition within functionalism of proposing inter-nal states that might lie behind regularities of behavior.Despite the efforts of Donald Hebb and Clark Hull to de-velop detailed process theories, the accepted approach instudies of human learning was to explore functional rela-tionships between variables and not to postulate hiddenprocesses. Even in 1969 this view was so strong that LyleBourne, a contemporary functionalist, in presenting researchon the formation of concepts, said, "There might not beanything missing in an account of behavior which fails tomention underlying psychological processes." This led AllenNewell to reply "to believe that we should proceed onlywith descriptions of regularities and avoid any attempt tosee in them the processing that is involved seems to mealmost a failure of nerve." If these powerful emotions (sinceI attended the meeting, I know they were) about the legiti-macy of inferring underlying process could be invoked in1969, how much more controversial was their use by Melton,a scion of the functionalist movement in 1940?

Melton and Irwin had subjects learn two lists of associ-ated word pairs, and then studied the recall of list oneassociations after varying amounts of list two learning. Theyplotted the amount of interference on list one memory asa function of the number of trials on list two. All of thiswas very consistent with the functionalist tradition of relat-ing an independent variable (number of list two trials) toa dependent variable (amount of retroactive interference).Then, however, Melton took a step beyond functionalisminto the internal world of process. The standard way tointerpret retroactive interference was that when subjectsattempted to produce associations to list one, they intrudedthe new learning developed during the list two trials. Meltonand Irwin measured the number of such overt intrusionsas a function of the number of list two learning trials.


Then they subtracted overt intrusions from the total amountof retroactive interference they had measured. What wasleft was a new hidden factor they called factor X. Not onlywas factor X large, but it also showed a systematic increasewith the amount of interpolated learning of list two. Atlow levels of list one learning, most of the interference wasdue to overt intrusions, but as list one learning increased,the interference appeared to be caused mainly by the in-ternal process called factor X. Melton tried to identifyfactor X with the unlearning of an association that hadbeen described by Clark Hull during extinction in classicalconditioning. Although the very label factor X indicates thecare with which Melton took to identify the theoretical natureof this hidden factor, it is clear that he was using the carefulfunctional approach to make visible a process that was other-wise hidden from view. The use of subtraction methods toreveal underlying process was to become a major part of newstudies of short-term memory and cognition that emergedafter World War II, but Melton's own research was a clearforerunner of the newer process orientation.

MILITARY PSYCHOLOGY

The combination of Melton's administrative skills as Psychol-ogy Department head at Missouri and his encyclopedic knowl-edge of psychological methods (as revealed by his paperson methodology) and applications (as revealed by his mu-seum studies) must have made him a nearly irresistibleattraction to the growing war effort by psychologists. Evenbefore Pearl Harbor he had worked on a battery of psycho-motor tests for pilot selection. Those tests were instru-mental in the smooth functioning of pilot selection afterthe war began. Melton wrote a long monograph on thosetest batteries. The success of those tests was important tothe prestige of experimental psychology in the military.


Melton's research operation was successful in perform-ing a valuable function in pilot selection during and follow-ing the war. It also served as a training ground for a largenumber of psychologists who were to become prominentleaders in the field of experimental psychology after thewar. Melton worked on military projects from 1941 until1957, with the exception of a brief period at Ohio (1946-49) that was interrupted for a one-year visit to Stanford in1948. Melton's administrative leadership helped to createand cement an important role for psychology within themilitary that has lasted until the present time.

HUMAN PERFORMANCE CENTER

Melton returned to academia at the University of Michi-gan in 1957. One year later he attracted his close col-league in military psychology, Paul Fitts, from Ohio StateUniversity. Together they built a major center for experi-mental psychology and its applications. Melton and Fittsregarded the field of human performance as the appropri-ate basic science out of which would flow applications tothe design of a wide range of civilian and military tech-nologies. Much current work in the design of technologythat can be used by humans flows from the concepts thatwere developed and the students who were trained at theHuman Performance Center. Human performance theoryemphasized the development of functional relationshipsthat might be explained in terms of transformations ofinternal representations. It became a part of the moregeneral cognitive psychology that developed in the 1960s.

Melton's unique research contribution was in the areaof short-term memory. This topic allowed Melton to com-bine his long-standing interest in verbal learning with thenewer information processing approach that emphasizedmemory within the context of cognitive tasks. Melton's


work stressed the continuity between short-term and long-term memory. His 1963 paper on short-term memory wasfar more consonant with current views of short-term memoryas an active state of long-term memory than it was with thethen popular buffer models. An example taken from thatpaper is the improvement in recall of a nine-digit numberwhen its repetitions were separated by two to eight othersequences. Even a single repetition of a nine-digit list wassufficient to improve its second recall after eight interven-ing lists. This could occur even when subjects had no ideathat lists were repeated. Melton concluded that "the struc-tural memory trace established by a single occurrence ofan event must be extraordinarily persistent." The evidencefor continuity between short- and long-term memory pro-vided by that paper was influential and remains widelycited as a source of information on the basic characteris-tics of short-term memory.

Another finding made by Melton (1967) supported theidea of continuity between short- and long-term memory.In studies of successive items within a single list, he foundthat the more widely spaced two occurrences were, themore likely an item was to be recalled on a subsequentoccasion. This effect was similar to the advantage of spacedpractice found in long-term memory skills. However, itwas possible to demonstrate the advantage of spaced rep-etition very quickly and reliably within the short-term mem-ory context. This made possible studies of the underlyingprocess that was at work to produce improved recall withincreased lag between repetitions (lag effect). One couldask, for example, whether the lag effect depended on themore varied context of presentation as the lag increased,whether there was reduced attention to the second itemwhen it followed immediately after the first, or whetherthe additional practice involved in reinstating a delay item


helped the subsequent recall. Once again Melton's explo-ration of a systematic function between two variables openedthe way for studies of the underlying psychological andphysiological mechanisms that are still continuing.

The Human Performance Center produced many of themajor figures in contemporary experimental psychology.Edwin Martin, Robert Bjork, and Judith Goggen were amongthe younger associates who began their academic careersby working with Melton in the research programs of thecenter.

A number of very important doctoral dissertations werealso developed there. Probably the most influential wasGerald Reicher's demonstration that every letter within aword is seen at a lower duration threshold than the sameletter when presented in isolation or within a nonsensestring. This finding was at first widely doubted, but it wasthen replicated innumerable times and later was a majorcontributor to the development of multilevel parallel pro-cessing models of cooperative computation that have beenso widely influential in recent years. Robert Crowder's veryconvincing demonstration that nonverbal interpolated mentaleffort reduces short-term memory for verbal items was abasic contribution to capacity models of short-term memory.Barbara Tversky's study of picture word matching were amongthe earlier objective experimental demonstrations of thedevelopment of a visual image from a verbal presentation.Harvey Shulman demonstrated the role of semantic codesin short-term retention, thus providing an additional dem-onstration of the continuity between short- and long-termretention. Richard Pew, Irving Biederman, Edward Smith,and Howard Egeth were among other graduates of the Center'sprogram during Melton's tenure.

Melton's contribution to the field of psychology includedtwelve years as editor of the Journal of Experimental Psychol-


ogy (1951-63). Melton was a believer in thorough experi-mental designs that explored multiple levels of a givenvariable and that used counterbalancing to avoid confound-ing influences. His years as editor helped to increase theconcerns of a generation of young investigators with theimportance of a high level of experimental control. Helater served as chief editorial adviser to the American Psy-chological Association. Honors were also numerous, in-cluding election to the National Academy of Sciences in1969 and receipt of the Gold Medal Award of the Ameri-can Psychological Association in 1976.

These contributions gave Art Melton a level of authoritywith the experimental psychologists of his era that is diffi-cult to describe. Melton's sense of dedication to psychol-ogy was so clear that one felt a greater debt because of it.It was a rare person who could refuse a request from Meltonto serve on a committee, review a paper, or perform someother service that Melton felt was needed by the field.

Melton's studies of retroactive interference, spacing, andshort-term memory helped to steer the field from the de-scriptive concerns he learned from the functionalists to aninterest in the underlying mental operations that mightserve as the empirical basis for a deeper theory of the hu-man mind. Even more than his individual studies, theenvironment for training that he worked to create at theUniversity of Missouri, in the Air Force, and most of all atthe University of Michigan provided several generations ofresearchers with the methods to assist in the development ofthis new process-oriented view of human abilities. AmongMelton's many awards was recognition for his graduate teachingfrom students at Michigan for his many contributions totheir education. Through their work as well as his own,Melton's ideas and values continue to contribute to ourunderstanding of psychology and to its application.


i WAS GREATLY ASSISTED by three more personal publishedbiographies concerned with Melton's work. One of thesewas an obituary written by Melton's long-time colleagueBenton Underwood (American Psychology 34[1979]:1171-73).Another was written jointly by Robert S. Daniel, Frank Geldard,James Greeno, and Wilert McKeachie (American Journal ofPsychology 95[1980]:153-58). Another biography appearedin conjunction with his reception of the Gold Medal Awardof the American Psychological Association (American Psy-chologist 32[1977]:98-99). Dr. Irwin Pollack, who has spentmany years at the University of Michigan, collected mate-rial that was of very substantial aid in developing the bibli-ography that appears following this memoir, and I am verygrateful to him. Mary K. Rothbart helped me to clarifythis memoir. Her editorial skill is greatly appreciated.


SELECTED BIBLIOGRAPHY

1929With J. A. McGeoch. The comparative retention values of maze

habits and nonsense syllables./. Exp. Psychol. 12:392—414.

1935Problems of Installation in Museums of Art. NS No. 14, pp. 1-269. Washing-

ton, D.C.: Publications of the American Association of Museums.

1936With N. G. Feldman, and C. W. Mason. Experimental Studies of the Edu-

cation of Children in Museums of Science. NS No. 15, pp. 1-106. Washington,D.C.: Publications of the American Association of Museums.

The methodology of experimental studies of human learning andretention. I. The functions of a methodology and the availablecriteria for evaluating different experimental methods. Psychol.Bull. 33:305-94.

The end-spurt in memorization curves as an artifact of the averag-ing of individual curves. Psychol. Monogr. 47:119-34.

1940With J. McQ. Irwin. The influence of degree of interpolated learn-

ing on retroactive inhibition and the overt transfer of specificresponses. Am. J. Psychol. 53:173-203.

1941With W. J. VonLackum. Retroactive inhibition and proactive inhibi-

tion in retention: Evidence for a two-factor theory of retroactiveinhibition. Am. J. Psychol. 54:157-73.

Learning. In Encyclopedia of Educational Research, ed. W. S. Munroe,pp. 667-86. New York: Macmillan.

The logic of modern psychology. I. Psychol. Bull. 38:227-40.

1944The selection of pilots by means of psychomoter tests. / . Aviat. Med.

15:116-23.

1947Apparatus Tests. AAF Aviation Psychology Research report no. 4,

Washington, D.C.: U.S. Government Printing Office.


1950Learning. Annu. Rev. Psychol. 1:9—30.

1956Present accomplishment and future trends in problem-solving and

learning theory. Am. Psychol. 11:278-81.

1957Military psychology in the United States of America. Am. Psychol.

12:740-46.

1959The science of learning and the technology of educational meth-

ods. Harv. Ed. Rev. 29:96-106.1960

With G. E. Briggs. Engineering psychology. Annu. Rev. Psychol. 11:71-89.

1963Implications of short-term memory for a general theory of memory.

/ Verb. Learn. Verb. Behav. 2:1-21.

1964Categories of Human Learning. New York: Academic Press.

1967Repetition and retrieval from memory. Science 158:232.

1970With E. Martin. Meaningfulness and trigram recognition. J. Verb.

Learn. Verb. Behav. 9:126-35.The situation with respect to the spacing of repetitions and memory.

/ . Verb. Learn. Verb. Behav. 9:596-606.Short and long-term postperceptual memory; dichotomy or con-

tinuum? In Biology of Memory, ed. K. H. Pribram, pp. 3-5. NewYork: Academic Press.

1972With E. Martin. Coding Processes in Human Memory. Washington, D.C.:

V. H. Winston and Sons.

1906—1978 · arthur weever melton august 13, 1906-november 5, 1978 by michael i. posner arthur...

Documents