MINNESOTA

In the spring of 1936, the low point of the depression, the end of my junior Fellowship was approaching and I had no job. The best offer the Department of Psychology could pass along to me was from a YMCA college; but Walter Hunter was teaching that summer at Minnesota, and he mentioned me to R.M. Elliott, who was looking for someone to teach small sections of a big introductory course. The beginning salary was $1900.

At Minnesota I not only taught for the first time, I began to learn college psychology, keeping a jump or two ahead of my students in Woodworth's text. I chose two sections of 20 students each from about 800 in the beginning course. Many of them were already committed to particular careers, such as medicine, law, journalism, and engineering, but 5 percent of the students I had during 5 years went on to get Ph.D.'s in psychology and many more to get M.A.'s. I stole W.K Estes from engineering and Norman Guttman from philosophy. My teaching has never again been so richly reinforced.

VERBAL BEHAVIOR

I did not quite give up literature. At Harvard I met IA Richards, who managed somehow to blend psychology and literary criticism, and I discussed books and techniques with other literary friends. I wrote an article for the Atlantic Monthly under the editor's title of "Has Gertrude Stein a Secret?" In it I showed that a paper which Gertrude Stein had published when at Radcliffe contained samples of her own automatic writing which resembled material she later published as literature. Gertrude Stein wrote to the editor in reply: "No, it is not so automatic as he thinks. if there is anything secret it is the other way too. I think I achieve by xtra consciousness, xcess, but then what is the use of telling him that, he being a psychologist and I having been one."

I began to look at literature, not as a medium for portraying human behavior, but as a field of behavior to be analyzed. A discussion with Whitehead after dinner at the Society of Fellows set me to work on my book Verbal Behavior (1957). The chairman of the Society, L.J. Henderson, cautioned me that such a book might take 5 years. The following summer he sent a postcard from France: "A motto for your book -- 'Car le mot, çest le verbe, et le verbe, ç'est Dieu -- Victor Hugo. "

As a boy I knew two interesting cases of verbal behavior. My Grandmother Skinner was an almost pathological talker. My grandfather had stopped listening to her while still a young man, and when any visitor came to her house she would begin talking and would repeat, without pausing, a string of anecdotes and stereotyped comments which we all knew by heart. More predictable verbal behavior I have never seen. The other case was Professor Bowles, the principal of my high school, who taught mathematics. He had a long list of favorite topics, and almost any stimulus would set him off on a digression. He would eventually return to mathematics with a perfunctory bow to the comment which had first set him off. One day I made running notes of the topics he was touching upon. There were two long harangues that day, and to my surprise he concluded the second by returning to the topic with which he had begun and concluded the first!

When I was in the Society of Fellows, another verbal phenomenon came to my attention. On a beautiful Sunday morning I was in my subterranean, soundproofed laboratory writing notes against a background of rhythmic noise from my apparatus. Suddenly I found myself joining in the rhythm, saying silently, "You'll never get out, you'll never get out, you'll never get out." The relevance of the remark seemed worth investigating. I built a phonographic system in which patterns of vowels (separated by glottal stops) could be repeated as often as desired. Playing each sample softly to a subject, I could maintain the illusion that it was actual speech and could collect a large sample of "projective" verbal responses. Harry Murray supplied me with subjects from his research on thematic apperception.

My renewed interest in literature was encouraged by my marriage in 1936 to Yvonne Blue. She had majored in English at the University of Chicago, where she had taken a course in English composition with Thornton Wilder. She is an active reader (and a rapid one -- she reads exactly twice as fast as I), and there were alewives new books around the house. When I had a chance to give a summer school course in the psychology of literature, she attended my lectures and reinforced me appropriately. I gave the course again and broadcast it over an educational radio station. To fill out the term I roamed rather widely, from The Meaning of Meaning (15) through psychoanalysis, and thus explored the field of verbal behavior rather more widely than I should otherwise have done. As a rule the material in which I had least confidence proved to be most popular, but I did not wholly abandon my scientific principles. After several persuasive demonstrations of alliteration as a verbal process, for example, I became suspicious and made a statistical analysis of 100 of Shakespeare's sonnets. I found that, although an occasional line might have as many as four stressed initial s's, such lines occurred almost exactly as often as one would predict from chance. (A similar study of Swinburne, I was glad to find, not only demonstrated alliteration, but showed an alliterative tendency extending over several syllables.)

In the fall of 1941 on a Guggenheim Fellowship I began to write a final draft of Verbal Bebavior. The war intervened, but I picked up the Fellowship again in 1944-45 and finished the greater part of the manuscript. I gave a course from it in the summer of 1947 at Columbia, and my William James Lectures at Harvard that fall were based on it. I put off a final version in order to write Science and Human Behavior (1953). Verbal Behavior was published, not 5, as Henderson had predicted, but 23 years after it was begun, in 1957. It was completed under heavy competition from research and from another book, Schedules of Reinforcement (1957), which Charles Ferster and I published at about the same time.

PROJECT PIGEON

By the end of the 1930s the Nazis had demonstrated the power of the airplane as an offensive weapon. On a train from Minneapolis to Chicago in the spring of 1939, 1 was speculating rather idly about surface-to-air missiles as a possible means of defense. How could they be controlled? I knew nothing about radar, of course, but infrared radiation from the exhaust of the engines seemed a possibility. Was visible radiation out of the question? I noticed a flock of birds flying alongside the train, and it suddenly occurred to me that I might have the answer in my own research. Why not teach animals to guide missiles? I went back to Minneapolis and bought some pigeons. The rest of the story of Project Pigeon has already been told. (16)

THE "BABY BOX"

Toward the end of the Second World War, we decided to have a second child. My wife remarked that she did not mind bearing children but that the first 2 years were hard to take. I suggested that we mechanize the care of a baby. There is nothing natural about a crib. Wrapping a baby in several layers of cloth -- undershirt, nightie, sheets, and blankets, with a mattress underneath -- is an inefficient way of maintaining a proper temperature, and it greatly restricts the child's movements. I built, instead, an enclosed space in which the baby, wearing only a diaper, could lie on a tightly stretched woven plastic sheet, the surface of which feels rather like linen and through which warm air rises, moved by convection or a fan, depending on the outside temperature.

When our second daughter, Deborah, came home from the hospital, she went directly into the device and used it as sleeping space for 2 1/2 years. I reported our happy experience in an article in the Ladies' Home Journal, and many hundreds of babies have been raised in what is now called an Aircrib. Child care is conservative, and the method has been adopted fairly slowly, but medical and behavioral advantages should be studied. Predictions and tales of dire consequences have not been supported. Deborah broke her leg in a skiing accident but presumably not because of "the box." Otherwise she has had remarkably good health. She is now in college, interested in art and music from Bach to Beattle, and she usually beats me at chess. To complete the story of the shoemaker's children, our older daughter, Julie, is married to a sociologist, Ernest Vargas, and is finishing her work for a Ph.D. in educational research. Their first child, Lisa, is, of course, being raised in an Aircrib. (17)

WALDEN TWO

In the spring of 1945 at a dinner party in Minneapolis, I sat next to a friend who had a son and a son-in-law in the South Pacific. I expressed regret that when the war was over they would come back and take up their old way of life, abandoning their present crusading spirit. She asked me what I would have them do instead, and I began to discuss an experimental attitude toward life. I said that some of the communities of the 19th century represented a healthy attitude. She pressed me for details and later insisted that I publish them. I was unaware that I was taking her seriously. A paper on "The Operational Analysis of Psychological Terms" (1945) was due on June 1, and I met that deadline. Then, to my surprise, I began to write Walden Two (1948). It began simply as a description of a feasible design for community living. I chose the unoriginal utopian strategy of having a few people visit a community. The characters soon took over.

In general I write very slowly and in longhand. It took me 2 minutes to write each word of my thesis and that is still about my rate. From 3 or 4 hours of writing each day I eventually salvage about 100 publishable words. Walden Two was an entirely different experience. I wrote it on the typewriter in 7 weeks. It is pretty obviously a venture in self therapy, in which I was struggling to reconcile two aspects of my own behavior represented by Burris and Frazier. Some of it was written with great emotion. The scene in Frazier's room, in which Frazier defends Walden Two while admitting that he himself is not a likeable person or fit for communal life, I worked out while walking the streets near our house in St. Paul. I came back and typed it out in white heat.

I receive a steady trickle of letters from people who have read Walden Two, want to know whether such a community has ever been established, and, if so, how they can join. At one time I seriously considered an actual experiment. It could be one of the most dramatic adventures in the 20th century. It needs a younger man, however, and I am unwilling to give up the opportunity to do other things which in the long run may well advance the principles of Walden Two more rapidly. A conference organized to consider an actual experiment was recently attended by nearly 100 people. (18)

INDIANA

In the fall of 1945 I became chairman of the Department of Psychology at Indiana. I took with me from Minnesota the unfinished manuscript of Verbal Bebavior, the manuscript of Walden Two, the Aircrib with its lovely occupant, and a miscellaneous lot of apparatus. I was inexperienced as an administrator, but the department survived my brief chairmanship. I did no undergraduate teaching, but the chapter in Science and Human Behavior on self control is to a large extent the joint product of a seminar in which, for almost the only time in my life, I successfully managed group thinking. In spite of my administrative responsibilities I ran a number of experiments -- all with pigeons -- on reaction time, differential reinforcement of slow responding, two operanda, and matching-to-sample. These studies are mostly reported in "Are Theories of Learning Necessary?" (1950).

THE EXPERIMENTAL ANALYSIS OF BEHAVIOR

Other people were now beginning to do research along the same lines. W.K. Estes, who went on to get a Ph.D. at Minnesota, wrote a thesis on the effects of punishment which became a classic. At Columbia Fred Keller was teaching graduate students from The Behavior of Organisms and, with W. N. Schoenfeld, was planning a revolutionary introductory course in the college. A problem in communication arose, and Keller and I started what became a series of annual conferences on the experimental analysis of behavior. Those who attended the first of these at Indiana in the spring of 1946 are pictured in volume 5 (1962) of the Journal of the Experimental Analysis of Behavior. Eventually we began to meet at the same time as the American Psychological Association and later as part of its program. When Division 3 could no longer provide space or arrange time for our expanding activities, we took the probably inevitable step of forming a separate division -- Division 25.

Meanwhile, the need for a special journal had become clear. I proposed an inexpensive newsletter, but more constructive opinions prevailed. A small holding society was formed and the Journal of the Experimental Analysis of Bebavior founded. The history of the discipline can also be traced in the increasing availability of excellent apparatus, reflecting the growing complexity and subtlety of the contingencies of reinforcement under analysis. (19)

HARVARD AGAIN

While giving the William James Lectures at Harvard in 1947, I was asked to become a permanent member of the department, and we moved to Cambridge in 1948. Remembering my introductory teaching at Minnesota, I proposed to add a course in human behavior to the Harvard list. The first year was nearly a disaster. More than 400 students, anticipating a "gut" course, signed up. I had no appropriate text and could only supply hastily prepared mimeographed sheets. My section men were loyal but puzzled. Later the course was incorporated into the General Education program and gradually improved. By 1953 Science and Human Behavior was available as a text.

Meanwhile I had set up a pigeon laboratory in which Charles Ferster and I worked very happily together for more than 5 years. It was the high point in my research history. Scarcely a week went by without some exciting discovery. Perhaps the behavior we dealt with most effectively was our own. Near the end of our collaboration we found ourselves with a vast quantity of unanalyzed and unpublished data, and we proceeded to design an environment in which we could scarcely help writing a book. In it we both worked as we had never worked before. In one spring term and one long hot summer we wrote a text and a glossary and prepared over 1000 figures, more than 900 of which were published.

The success of my laboratory in the 1950s and early 1960s was due in large part to many excellent graduate students, not all of them under my, direction, of whom I may mention Douglas G. Anger, James E. Anliker, Donald S. Blough, Richard J. Herrnstein (now my colleague on the Harvard faculty), Alfredo V. Lagmay, William H. Morse, Nathan H. Azrin, Ogden R. Lindsley, Lewis R. Gollub, Matthew L. Israel, Harlan L. Lane, George S. Reynolds, A. Charles Catania, Herbert S. Terrace, and Neil J. Peterson. With very little direct help from me they all made and are continuing to make important contributions.

TECHNOLOGICAL APPLICATIONS

At Minnesota W.T. Heron and I had studied the effects of certain drugs on operant behavior. In the 1950s a strong interest in psychopharmacology suddenly developed. Almost all the large drug companies set up operant laboratories, some only for the screening of new compounds but many providing an opportunity for basic research. Much of this interest was generated by Joseph V. Brady of the Walter Reed Army Medical Center. Peter Dews of the Department of Pharmacology in the Harvard Medical School began to work in close cooperation with my laboratory and soon organized an active program in his own department.

In the early 1950s Dr. Harry Solomon, then chairman of the Department of Psychiatry at the Harvard Medical School, helped me set up a laboratory for the study of the operant behavior of psychotics at the Metropolitan State Hospital in Waltham, Massachusetts. Ogden R. Lindsley took over, and the work he initiated there has now been carried forward in many other laboratories. Azrin and others have extended operant principles to the management of psychotic patients in hospital wards, and there is increasing interest in applications to personal therapy. (20)

Sporadic research on operant behavior in children goes back to the 1930s. Sidney Bijou, among others, has been particularly active in applying the principles of an experimental analysis to the behavior of children in nursery schools, clinics, and the home. Ferster turned from our work on schedules to the study of autistic children, and there are now many operant laboratories for the study of retardates. Almost all these practical applications have contributed to our understanding of behavior. Fortunately, they have not overshadowed the basic science; many laboratories continue to study operant behavior apart from technological significances.

In the late 1930s, looking ahead to the education of our first child, I began to write a book called Something to Think About. It was never completed, though I got as far as having an artist work on the illustrations. It contained examples of what later came to be called programmed instruction. When our daughters went to school, I showed the usual interest as a parent but carefully refrained from speaking as a specialist in the field of learning. In 1953 our younger daughter was in fourth grade in a private school in Cambridge. On November 11, as a Visiting Father, I was sitting in the back of the room in an arithmetic class. Suddenly the situation seemed perfectly absurd. Here were 20 extremely valuable organisms. Through no fault of her own the teacher was violating almost everything we knew about the learning process.

I began to analyze the contingencies of reinforcement which might be useful in teaching school subjects and designed a series of teaching machines which would permit the teacher to provide such contingencies for individual students. At a conference on current trends in psychology at the University of Pittsburgh in the spring of 1954 I demonstrated a machine to teach spelling and arithmetic. Within a year I found myself caught up in the teaching machine movement. A series of projects at Harvard led eventually to a Committee on Programmed Instruction, in which I had the invaluable collaboration of James G. Holland. (21)

Economics, government, and religion are farther from psychology than linguistics, psychotherapy, or education, and few people have the kind of joint interest needed for an examination of common principles. I have seen myself moving slowly in this direction, however, and I am now working under a Career Award from the National Institutes of Health which will permit me to explore the social sciences from the point of view of an experimental analysis of behavior . (22)