Bowlby and Harry Harlow on Attachment Behavior
Frank C. P. van der Horst1 , Helen A. LeRoy2 and René van der Veer1
(1) Centre for Child and Family Studies, Leiden University, P.O. Box 9555, NL-2300RB Leiden, The Netherlands
(2) Harlow Primate Laboratory, University of Wisconsin, Madison, WI, USA
Frank C. P. van der Horst
Email: fhorst@fsw.leidenuniv.nl
Received: 12 February 2008 Accepted: 21 August 2008 Published online: 3 September 2008
Abstract From 1957 through the mid-1970s, John Bowlby, one of the founders of attachment theory, was in close personal and scientific contact with Harry Harlow. In constructing his new theory on the nature of the bond between children and their caregivers, Bowlby profited highly from Harlow’s experimental work with rhesus monkeys. Harlow in his turn was influenced and inspired by Bowlby’s new thinking. On the basis of the correspondence between Harlow and Bowlby, their mutual participation in scientific meetings, archival materials, and an analysis of their scholarly writings, both the personal relationship between John Bowlby and Harry Harlow and the cross-fertilization of their work are described.
Keywords Attachment theory - Animal psychology - Ethology - Animal behavior - Infant–mother relations - History
Frank C.P. van der Horst is a PhD student and Lecturer at the Centre for Child and Family Studies at Leiden University, The Netherlands. The work presented in this special issue is part of his doctoral thesis on the roots of Bowlby’s attachment theory. The defence of this thesis, titled John Bowlby and ethology: a study of cross-fertilization, is scheduled for early 2009.
Helen A. LeRoy recently retired from the Harlow Primate Lab at the University of Wisconsin-Madison after working there for nearly half a century. During that time, she worked closely with Harry Harlow from her arrival in 1958 until his retirement in 1974. She was Harlow’s executive assistant and was his help and stay in the editing of the Journal of Comparative and Physiological Psychology.
René van der Veer is Professor of History of Educational Thinking at Leiden University, The Netherlands. His research addresses the work of key educational thinkers such as Gal’perin, Janet, Piaget, Vygotsky, Werner, and Wallon. In a longer study the origin of the idea of the social mind was traced. He is on the Editorial Board of Integrative Psychological and Behavioral Sciences.
Introduction
Today, one can pick up almost any introductory, general, or developmental psychology textbook (e.g., DeHart et al. 2004; Cole and Cole 2005) and find references to British child psychiatrist John Bowlby (1907–1990) and American animal psychologist Harry Harlow (1905–1981). Quite often their work is discussed in tandem. Bowlby was a clinician by training and Harlow an experimentalist. Despite these rather different backgrounds, the two men had several things in common. One of them was that they showed no hesitation in expressing views that went against the prevailing Zeitgeist. In the 1950s and 1960s, both Bowlby and Harlow formulated new ideas on the nature of the bond between child and caregiver. They defied the prevailing psychoanalytic and learning theoretical views that dominated psychological thinking from the 1930s. Although it has been argued (Singer 1975) that Harlow’s experimenting had no influence on Bowlby’s theorizing, here it will become clear that Bowlby used Harlow’s surrogate work with rhesus monkeys as much needed empirical support for his emerging theory of attachment in which he explained the nature and function of the affectional bonds between children and their caregivers (Bowlby 1958, 1969/1982). In his turn, Harlow was influenced by Bowlby’s thinking and tried to model his rhesus work to support Bowlby’s new theoretical framework (e.g., Seay et al. 1962; Seay and Harlow 1965).
The theories of Harlow and Bowlby are well-known but so far little was known about the personal and professional relationships between these two giants in the field. In this contribution, on the basis of the correspondence between Harlow and Bowlby1, their joint participation in scientific meetings, archival materials, and an analysis of their scholarly writings, an attempt is made to delineate the cross-fertilization of their work during the most active years of their acquaintance from 1957 through the mid-1970s. It will be demonstrated that Bowlby and Harlow’s interests converged as Harlow shifted his focus to a developmental approach shortly before the two met. Their introduction at a distance by British ethologist Robert Hinde was the beginning of an exchange of ideas that resulted in groundbreaking experimenting and theorizing that affects the field of developmental psychology to this day.
Bowlby’s Early Career (1938–1957): from Kleinian Psychoanalysis to Real Life
John Bowlby, who received a Master’s degree from Cambridge University and an M.D. from University College Hospital in London, was trained in psychoanalysis. He practiced as a clinician and joined the staff of the Tavistock Clinic in London in 1946, where he spent the remainder of his professional career (cf. Van Dijken 1998; Van Dijken et al. 1998). There is no doubt that he will be remembered in history as “the father of attachment theory”. Bowlby’s career evolved on the basis of a single theme, the relationship between mother and infant, and the effects of the pattern established early on upon the developmental outcome of the offspring. He mounted a scientific challenge to dominant psychoanalytical views in British psychiatry, such as those held by Anna Freud and Melanie Klein (Berrios and Freeman 1991).
In an interview with Robert Karen (1994, pp. 45–46), Bowlby described an influential experience in 1938, while training under the supervision of psychoanalyst Melanie Klein. Contrary to Klein, who believed all behavior was motivated by inner feelings, Bowlby felt that external relationships, e.g., the way a parent treated a child, were important to consider in understanding the child’s behavior. At the time, he was seeing an anxious, hyperactive child as a patient five days a week. The boy’s mother would sit in the waiting room, and Bowlby noticed that she too seemed quite anxious and unhappy. When he told Klein he wanted to talk to the mother as well, Klein refused adamantly, dismissing the mother as a possible causal or related factor in the child’s behavior. Bowlby was thoroughly annoyed and gradually distanced himself from the Melanie Klein school of thought. Later, in 1948, through the work of Tavistock social worker James Robertson, with whom he would work closely over the years, Bowlby became interested in recording the responses shown by children between the ages of 12 months and 4 years upon separation from their mothers or attachment figures (Bowlby 1960a).
In 1950, as part of a World Health Organization (WHO) project, Bowlby (1951) undertook a literature survey in order to test the hypothesis that “separation experiences are pathogenic” (Bowlby 1958). Homeless children had become a major problem after World War II, and in his WHO report, Bowlby warned that children deprived of their mothers were at risk for physical and mental illness. After surveying the literature, Bowlby et al. (1952, p. 82) concluded:
It became clear that this hypothesis is well supported by evidence and the team is now planning to concentrate on understanding the psychological processes which lead to the grave personality disturbances—severe anxiety conditions and psychopathic personality—which we now know sometimes follow experiences of separation.
We may conclude, then, that Bowlby was convinced at the time that (repeated) separation experiences may seriously harm the mental health of children and that the existing literature (e.g. on hospitalization) proved his point of view. He valued empirical studies and emphasized the importance of objective observation of real-life experiences. However, he still lacked the theoretical apparatus to understand the causal mechanisms behind the phenomena he observed. Also, he knew of no experiments that manipulated the potentially relevant variables in the domain of attachment formation. It was in this situation that he chanced upon the emerging science of ethology and the experimental work of Harlow.
In the subsequent years Bowlby made increasing use of ethological findings and theorizing guided by British ethologist, colleague and life-long friend Robert Hinde (Van der Horst et al. 2007). Bowlby (1957, 1960c) acknowledged a deep and pervasive interest in ethology beginning about 1951, which was sparked by Konrad Lorenz’s (1935, 1937) gosling work. His talk to the members of the British Psycho-Analytical Society on June 19, 1957 (published as Bowlby 1958) testifies of his growing confidence in the relevance of ethology.
Harlow’s Early Career (1930–1957): from Conditioning Rats to Studying Monkey Love
Harry Harlow received a Ph.D. in psychology from Stanford University in 1930 and spent the remainder of his academic career as a professor at the University of Wisconsin—Madison. Harlow was educated in the psychological tradition of the 1920s and 1930s, a time when psychology was making an effort to become a ‘real’ science. Studying behavior was a case of controlling the environment and varying one particular condition. It was a time when behaviorist views carried the day and the conditioned responses of Norwegian rats were the key to understanding mental life. So, when Harlow was appointed at Wisconsin in 1930 and found that the psychology department’s chairman had the rat laboratory dismantled and it was not about to be replaced, he was greatly inconvenienced (Harlow 1977, p. 138–139; Suomi and LeRoy 1982).
It was only at the suggestion of the chairman’s wife that Harlow decided to study primates at the local zoo and he soon found out that the intellectual capabilities of the monkeys were far greater than those of rats (Suomi and LeRoy 1982). To study these capabilities more rigorously and effectively Harlow developed the Wisconsin General Test Apparatus (WGTA; Harlow and Bromer 1938) by which it was possible to present the monkeys with a large number of learning tests in a highly standardized way. With it he tested the monkeys with discrimination learning and memory tasks (e.g., Harlow 1943, 1944). Harlow’s next step was to study cortical localization of learning capabilities by doing lesion studies with monkeys (e.g., Harlow and Dagnon 1943; Harlow and Settlage 1947; Moss and Harlow 1948). By lesioning different areas of the brain, Harlow noted that each of the operated monkeys performed differently on the WGTA tests. This work was basically similar to the work done by Lashley (e.g., Lashley 1950).
In the late 1940s, Harlow achieved “a major conceptual and methodological breakthrough” (Suomi and LeRoy 1982, p. 321) by identifying the formation of learning sets in monkeys (Harlow 1949). Harlow demonstrated that his monkeys “learned to learn” and that they acquired a strategy for problem-solving. As methods of studying processes underlying monkey learning were exhausted, Harlow in the early 1950s turned to studying motivation and the ontogeny of learning. This type of developmental research required the establishment of a breeding colony of rhesus monkeys. It was at this point that Harlow’s attention was drawn to the phenomenon of affection.
Harlow had always had problems importing monkeys: apart from being very expensive, they were often ill upon arrival and infected the other monkeys in the laboratory. In 1956, following procedures of Van Wagenen (1950), he decided to raise his own rhesus monkeys, and thus the Wisconsin lab became the first self-sustaining colony of monkeys in the US. The monkeys were kept separated at all times to avoid any spread of disease. The results of this procedure were remarkable for those who could see it: the monkeys Harlow raised were physically perfectly healthy, but their social behavior was very awkward. They were simply unable to socially interact with each other. Another striking observation Harlow made was that the infant monkeys “clung to [the diapers on the floor of their cage] and engaged in violent temper tantrums when the pads were removed and replaced for sanitary reasons” (Harlow 1958, p. 675). Harlow wondered whether these observations could mean anything for the needs of human children.
Just two months prior to Bowlby’s British Psycho-Analytical Society address which discussed in great depth the child’s tie to the mother, Harlow spoke on April 20, 1957 at a conference in Washington, D.C. The title of his address was the “Experimental Analysis of Behavior” and it included a discussion of trends in this area. Harlow began his address by stating that “no behavior is too complicated to analyze experimentally, if only the proper techniques can be discovered and developed” (Harlow 1957, p. 485). He went on to emphasize the importance of a developing trend toward longitudinal studies (psychology had traditionally been concerned with a cross-sectional approach), and he told how:
I have followed with interest the changes in my own research programs and the development of these programs. The experimental S that has consumed almost all my research time has been the rhesus monkey. When I initially approached the experimental analysis of this animal’s behavior, I approached it in the classical, cross-sectional manner (...). If it had not been for the fact that my monkey Ss continued to live after they had solved a problem and that they were not expendable in view of the available financial support, I might still be engaged in cross-sectional studies of the monkey’s behavior. (Harlow 1957, p. 487)
These comments clearly indicate that Harlow was moving towards experimental developmental research, the type of research that Bowlby so badly needed at the time. Harlow was now on the threshold of the affectional studies, for which he would become famous. He explained that:
More recently we have planned and initiated much more extensive longitudinal studies in which we have separated infant rhesus monkeys from their mothers at birth and raised them under the controlled conditions of the laboratory. We have been successful in raising over fifty of these young animals, and we have obtained data on their learning development from birth through three years of age. (...) We have found the longitudinal approach to the experimental analysis of behavior interesting and even exciting, and we are now extending this type of analysis to other areas than learning, perception, and motivation. (...) [W]e are planning and conducting systematic longitudinal studies on the development of emotional responses. (Harlow 1957, p. 488)
Just like Bowlby before his fellow psychiatrists of the British Psycho-Analytical society, Harlow (1957, p. 490), before an audience of clinical psychologists, stressed the importance of observational methods in this process, something that was of course very obvious to him.
At the present time (...) we are interested in tracing the development of various patterns of emotional behavior. (...) We began by looking for response patterns which might fit. (...) But this observational study (...) is gradually taking on the characteristics of an experiment. As we gain sophistication about the monkey’s emotional responses, we become more selective in the patterns which we observe.
It was because of their mutual interest in this area of emotional behavior and responses that Harlow and Bowlby became acquainted. In Harlow’s words: “It is an understatement to add that we have research interests in common” (Harlow in a letter to Bowlby, January 27, 1958).
Ethology and Animal Psychology: Contrasting Approaches to Animal Behavior
It was not self-evident for a British ethologically oriented psychiatrist and an American animal behaviorist to meet in those days. In the 1950s, there was a great barrier between ethologists (who were mostly biologists by training) and students of animal behavior (mostly psychologists). Ethologists emphasized observation of animals in their natural habitat, whereas comparative psychologists relied on rigorous experimentation in the laboratory. The culmination of this debate was a 1953 critique by Theodore Schneirla’s student Daniel Lehrman of Lorenz’s concept of instinct, at that time the central theoretical construct of ethology (Lehrman 1953). But in contrast to what might be expected, when Lehrman visited Europe in 1954 and met with leading ethologists, he was very well received. Just like many of the ethologists, Lehrman had a background in evolutionary biology and ornithology and this may have been essential in bridging their differences. Although Lorenz never acknowledged Lehrman’s ideas, they later became mainstream ethology (Griffiths 2004). Eventually Hinde (1966) wrote his authoritative book Animal behaviour which was essentially “a synthesis of ethology and comparative psychology” (cf. Van der Horst et al. 2007, p. 9–10).
In this climate of contrasting views, Hinde and Harlow met for the first time in Palo Alto in early 1957 at a conference organized by Frank Beach that was intended to bring together a group of European ethologists (Niko Tinbergen, Gerard Baerends, Jan van Iersel, David Vowles, Eckhardt Hess and Robert Hinde) with a group of mainly North-American comparative and experimental psychologists (Frank Beach, Donald Hebb, Daniel Lehrman, Jay Rosenblatt, Karl Lashley and Harry Harlow). Hinde has good memories of the event: “It was a wonderful conference, about three weeks, [where you had] nothing to publish, and if you did not finish what you had to say today there was always tomorrow” (Robert Hinde, personal communication, March 14, 2007). After their first encounter, Hinde and Harlow met several times in the late fifties and sixties. Although they influenced each other and their relationship was very cordial in the days they interacted, Hinde in retrospect remembers that at that time their approaches were still rather far apart:
I must have next met Harry when I visited Madison and was appalled by this room full of cages with babies going “whoowhoowhoo” and Harlow had no sensitivity at that point that he was damaging these infants. At that time I was beginning to work on mother–infant relations in monkeys myself, but I already knew enough about monkeys to know that that “whoo”-call was a distress call. These experiments had their restrictions, but they did show certain important things. After that I saw him at least once a year for a while as he asked me to join his scientific committee. Of course his results influenced my way of thinking, but I was then an ethologist and not keen on his laboratory orientation. And I could never have attempted to do the sort of research that he did because our colony only had six adult males and two or three females in each group. We attempted to create an approximation to a normal social situation: it was a long way off, of course, but at least it was social. (Robert Hinde, personal communication, August 22 and 26, 2005; March 14, 2007)
Despite these differences in theoretical orientation, it was Robert Hinde who would eventually establish contact between Bowlby and Harlow. At the Palo Alto conference, Hinde and Harlow had a discussion on motherhood and after returning home Hinde informed Bowlby that Harlow was interested in Bowlby’s recent work on this subject (Stephen Suomi, personal communication, September 27, 2006; Karen 1994; Hrdy 1999; Blum 2002; Van der Horst et al. 2007).
Harlow and Bowlby Become Acquainted in 1957
It was just several months later that Bowlby and Harlow introduced themselves by letter. The written record of their relationship commenced with a letter dated August 8, 1957 in which Bowlby expressed his interest:
Robert Hinde tells me that you were interested in my recent paper when he showed it to you at Palo Alto and at his suggestion I am now sending you a copy. I need hardly say I would be most grateful for any comments and criticisms you cared to make. I shall be at the Center at Palo Alto from mid-September and will be preparing it for publication then. Robert Hinde told me of your experimental work on maternal responses in monkeys. If you have any papers or typescripts I would be very grateful for them. If there were a chance, I would try to visit you next Spring when I hope to be moving around U.S.A. (Bowlby in a letter to Harlow, August 8, 1957)
The paper which Bowlby sent to Harlow at the time was a draft of “The nature of the child’s tie to his mother” (Bowlby 1958). Harlow replied by return of post, thanking Bowlby for the paper, which he several years later (in a letter to Bowlby, March 25, 1959) would refer to as a “reference bible”:
[Y]our interests are (...) closely akin to a research program I am developing on maternal responses in monkeys. I certainly hope that you can pay a visit to my laboratory sometime during this forthcoming year. At the moment our researches are just getting underway, and I hope to use these materials for my American Psychological Association Presidential Address in September, 1958. This address will be the first formal presentation of these researches. (Harlow in a letter to Bowlby, August 16, 1957)
Mutual Referencing After 1957
It was only after the two men began corresponding in August, 1957, that they began referring to each other’s writings. A review of Bowlby’s publications from 1951–1957 (Bowlby 1951, 1953, 1957; Bowlby et al. 1952, 1956; Robertson and Bowlby 1953) yields no mention of Harlow’s work. Likewise, we find no reference to Bowlby’s work in Harlow’s first developmental writings (Harlow 1957).