When psychoanalysis emerged from Vienna in the first decades of the twentieth century, it represented a revolutionary way of thinking about mind and its disorders. The excitement surrounding the theory of
unconscious mental processes increased as the century reached its midpoint and psychoanalysis was brought to the United States by émigrés from Germany and Austria.
But by 1960, when I began clinical training in psychiatry, my enthusiasm had stalled. My marriage to Denise, an empirical sociologist, and my research experiences—first in Harry Grundfest's laboratory at Columbia and then in Wade Marshall's laboratory at the National Institute of Mental Health—tempered my enthusiasm for psychoanalysis. While I still admired the rich, nuanced view of mind that psychoanalysis had introduced, I was disappointed during my clinical training to see how little progress psychoanalysis had made toward becoming empirical, toward testing its ideas. I also was disappointed in many of my teachers at Harvard, physicians who were motivated to enter psychoanalytic psychiatry out of humanistic concerns, as I was, but who had little interest in science. I sensed that psychoanalysis was moving backward into an unscientific phase and, in the process, was taking psychiatry with it.
Under the influence of psychoanalysis, psychiatry was transformed in the decades following World War II from an experimental medical discipline closely related to neurology into a nonem-pirical specialty focused on the art of psychotherapy. In the 1950s academic psychiatry abandoned some of its roots in biology and experimental medicine and gradually became a therapeutic discipline based on psychoanalytic theories. As such, it was strangely unconcerned with empirical evidence or with the brain as the organ of mental activity. In contrast, medicine evolved during this period from a therapeutic art into a therapeutic science, based on a reductionist approach derived first from biochemistry and later from molecular biology. During medical school, I had witnessed and been influenced by this evolution. I therefore could not help but note the peculiar position of psychiatry within medicine.
Psychoanalysis had introduced a new method of examining the mental life of patients, a method based on free association and interpretation. Freud taught psychiatrists to listen carefully to patients and to do so in new ways. He emphasized a sensitivity to both the latent and the manifest meaning of the patient's communications. He also created a provisional schema for interpreting what might otherwise appear as unrelated and incoherent reports.
So novel and powerful was this approach that for many years not only Freud but other intelligent and creative psychoanalysts as well could argue that psychotherapeutic encounters between patient and analyst provided the best context for scientific inquiry into mind, particularly into unconscious mental processes. Indeed, in the early years psychoanalysts made many useful and original observations that contributed to our understanding of mind simply by listening carefully to their patients and by testing the ideas that arose from psychoanalysis—such as childhood sexuality—in observational studies of normal child development. Other original contributions included the discovery of different types of unconscious and preconscious mental processes, the complexities of motivation, transference (the displacing of past relationships onto the patient's current life), and resistance (the unconscious tendency to oppose a therapist's efforts to effect change in the patient's behavior).
Sixty years after its introduction, however, psychoanalysis had exhausted much of its novel investigative power. By 1960 it was clear, even to me, that little in the way of new knowledge or insights remained to be learned by observing individual patients and listening carefully to them. Although psychoanalysis had historically been scientific in its ambitions—it had always wanted to develop an empirical, testable science of mind—it was rarely scientific in its methods. It had failed over the years to submit its assumptions to replicable experimentation. Indeed, it was traditionally far better at generating ideas than at testing them. As a result, psychoanalysis had not made the same progress as some other areas of psychology and medicine. Indeed, it seemed to me that psychoanalysis was losing its way. Rather than focusing in on areas that could be tested empirically, psychoanalysis expanded its scope, taking on mental and physical disorders that it was not optimally suited to treat.
Initially, psychoanalysis was used to treat what were called neurotic illnesses: phobias, obsessional disorders, and hysterical and anxiety states. However, psychoanalytic therapy gradually extended its reach to almost all mental illnesses, including schizophrenia and depression.
By the late 1940s, many psychiatrists, influenced in part by their successful treatment of soldiers who had developed psychiatric problems in battle, had come to believe that psychoanalytic insights might be useful in treating medical illnesses that did not respond readily to drugs. Diseases such as hypertension, asthma, gastric ulcers, and ulcerative colitis were thought to be psychosomatic—that is, induced by unconscious conflicts. Thus by 1960 psychoanalytical theory had become for many psychiatrists, particularly those on the East and West coasts of the United States, the prevailing model for understanding all mental and some physical illnesses.
This expanded therapeutic scope appeared on the surface to strengthen psychoanalysis's explanatory power and clinical insight, but in reality it weakened psychiatry's effectiveness and hindered its attempt to become an empirical discipline aligned with biology. When Freud first explored the role of unconscious mental processes in behavior in 1894, he was also engaged in an effort to develop an empirical psychology. He tried to work out a neural model of behavior, but because of the immaturity of brain science at the time, he abandoned the biological model for one based on verbal reports of subjective experiences. By the time I arrived at Harvard to train in psychiatry, biology had begun to make important inroads in understanding higher mental processes. Despite these advances, a number of psychoanalysts took a far more radical stance—biology, they argued, is irrelevant to psychoanalysis.
This indifference to, if not disdain for, biology was one of the two problems I encountered during my residency training. An even more serious problem was the lack of concern among psychoanalysts for conducting objective studies, or even for controlling investigator bias. Other branches of medicine controlled bias by means of blind experiments, in which the investigator does not know which patients are receiving the treatment being tested and which ones are not. However, the data gathered in psychoanalytic sessions are almost always private. The patient's comments, associations, silences, postures, movements, and other behaviors are privileged. Of course, privacy is central to the trust that must be earned by the analyst—and therein lies the rub. In almost every case, the only record is the analyst's subjective accounts of what he or she believes happened. As research psychoanalyst Hartvig Dahl has long argued, such interpretation is not accepted as evidence in most scientific contexts. Psychoanalysts, however, are rarely concerned about the fact that accounts of therapy sessions are necessarily subjective.