Did Freud's Psychoanalysis Work? Anna O., the Talking Cure and Eysenck's Count, 1880–1952

This is the third part of a series on the history of psychiatry. Part 2 ended with Kraepelin sorting mental illnesses of unknown cause by their course. In Vienna, in the same years, another physician took a completely different road. Sigmund Freud looked for causes not in the brain but in the memories his patients told him. His method, psychoanalysis, went on to dominate American psychiatry by the middle of the twentieth century. Did it work? This part looks at the case it started from, and at a recount made in 1952.

This series is a history of psychiatry. It is not advice on diagnosis or treatment. If you are struggling with your mental or physical health, please talk to a doctor or a support service. The text includes names of illnesses and expressions that are no longer in use; they are given as they appear in the sources of the time.

Patients with nothing to find in the brain

At the end of the nineteenth century one illness was a standing embarrassment to physicians: hysteria. A limb was paralysed, sight failed, the voice went. Yet on examination there was no injury or lesion to account for it. In this article the word is used only as the diagnosis of the time.

In a lecture of 1909 Freud described how doctors reacted. Faced with a disease of the brain, a physician can picture what is going on inside the head. Faced with hysteria, all his anatomy and physiology desert him, and he stands where the layman stands. Nobody who values his knowledge enjoys that. So the physician blames the patient for exaggeration and pretence, and punishes her by withdrawing his interest.

General paresis, in Part 2, was the illness whose cause was found in the brain. Hysteria sat at the opposite pole.

1880: Breuer's patient "Anna O." and the talking cure

Freud opened that lecture with an unexpected remark. If it is a merit to have created psychoanalysis, he said, it is not his merit. He was still a student, busy with his last examinations, when another Viennese physician, Josef Breuer, first applied the method to a patient. That was in 1880 to 1882.

The patient was a woman of twenty-one, of high intelligence. In the book published later she is called "Anna O." Her symptoms began while she was nursing her gravely ill father: paralysis of the right arm and leg, disturbed eye movements, a severe cough. For a time she could neither speak nor understand German, her mother tongue, and talked only in English.

Breuer noticed the words she muttered in her absent states. He put her under hypnosis and used those words as a starting point. She then told him, one after another, the daydreams that had been occupying her, and for a while afterwards she was calm. She herself named the procedure, in English, the "talking cure," and jokingly called it "chimney sweeping."

One episode became famous. In a hot summer she found herself unable to drink for several weeks. She would lift a glass and push it away as it touched her lips. One day, under hypnosis, she described with disgust how she had seen the little dog of her English governess, whom she disliked, drink out of a glass. At the time she had said nothing, out of politeness. Once she had given vent to her anger she asked for water and drank a large amount. The symptom, Breuer wrote, never came back.

A symptom, on this view, is the residue of an experience charged with a feeling that could not be expressed. Recall the scene, let the feeling out, and the symptom goes. Freud put it in one sentence: hysterical patients suffer from reminiscences.

What happened after the "cure"

The case has long been told as the starting point of psychoanalysis. But the illness went on far longer than the usual telling suggests.

Breuer's own wording was guarded. In Studies on Hysteria, the book he published with Freud in 1895, he closed the case by saying that she left Vienna and travelled for a while, but that it was a considerable time before she regained her mental balance entirely. Since then, he added, she had enjoyed complete health.

In 1972 the historian of psychiatry Henri Ellenberger published a critical review of the case with new data. Drawing on the follow-up literature, the Canadian psychiatrist Harold Merskey summed up the course in 1992. The illness was "a very protracted one," with fluctuations or exacerbations lasting from 1880 to 1887. It was complicated by dependence on morphine and on chloral hydrate, a sleeping drug. Only by 1888 does she appear to have made a considerable recovery.

Breuer's treatment ended in 1882. Recovery, then, took about six more years. Merskey also notes that she went on to lead an effective and fruitful life. Her real name was Bertha Pappenheim.

So this much is known. There is a record of individual symptoms disappearing after she talked. It cannot be said that the illness itself was cured by the method. And whether her eventual recovery was due to treatment or to the passage of time cannot be decided from a single case.

1896: eighteen patients and "a scientific fairy tale"

Freud went on to use Breuer's method on his own patients, modifying it as he went. In the spring of 1896 he lectured on the cause of hysteria to the society for psychiatry and neurology in Vienna. His letters point to 21 April as the date.

The claim was a strong one. In all eighteen patients he had analysed, six men and twelve women, the chain of memories led back to sexual abuse suffered in earliest childhood. At the bottom of every case of hysteria, he said, there were experiences of this kind. Most of the analyses, he noted, had taken a hundred hours or more.

The reception was cold. According to a letter Freud sent to a friend, the chairman, Richard von Krafft-Ebing, said it sounded like "a scientific fairy tale." He is the same Viennese physician who appeared in Part 2.

What is striking is that Freud raised the most damaging objection himself. Might the physician be forcing such scenes on docile patients? Might the patients be inventing them, and the physician believing what he is told? Freud replied that he had never yet succeeded in forcing on a patient a scene he was expecting to find.

Yet Freud later withdrew the theory. In a footnote added to the same paper in 1924 he wrote that at the time he had not yet freed himself from his overvaluation of reality and his low valuation of phantasy. The theory swung round: the scenes patients described came to be read not as events but as fantasies and wishes.

This article does not judge whether those patients' experiences were real. The point here is the method of checking. The first theory and the one that replaced it rested on the same material: what patients said in the consulting room, and how Freud read it. In neither case was there a way for someone else to test the claim by other means.

Popper's objection: a theory that explains everything

The philosopher Karl Popper stated this weakness most sharply. In a lecture of 1953 he looked back on doubts he had formed in Vienna around 1919.

Popper set two actions side by side. One man pushes a child into the water, meaning to drown it. Another gives his life trying to save the child. On Freud's theory the first man suffers from repression, and the second has turned the same impulses into something admirable. Opposite actions, and the theory explains both.

His conclusion: there was "no conceivable human behaviour" that could contradict such theories. That looks like their strength and is in fact their weakness, because a theory that cannot fail cannot be tested. Popper did grant that psychoanalysis contains "most interesting psychological suggestions." His complaint was that they were not in a testable form.

The criticism has been answered. According to the Stanford Encyclopedia of Philosophy, the philosopher of science Adolf Grünbaum objected that Popper's example is contrived and not taken from Freud's actual clinical texts. And it is true that the 1896 claim, "in all eighteen cases," was a strong one that could have failed. Freud himself gave it up.

1941 and 1952: the same kind of numbers, opposite conclusions

Psychoanalysis spread all the same, above all in the United States. Freud's 1909 lectures were given at Clark University in Worcester, Massachusetts, the town whose state hospital figures Pliny Earle recounted in Part 1. According to a 2010 paper by Kenneth Kendler and colleagues, psychoanalysis came to dominate American psychiatry in the 1940s, and by the mid-1950s nearly every department chairman of psychiatry in the country was an advocate of it.

Some tried to show its effect in numbers. In 1941 the psychoanalyst Robert Knight tabulated 952 cases in which analysis had been attempted, by diagnosis and result. He concluded that psychoanalysis "must be adjudged an effective therapy" for the neuroses. Neurosis was the catch-all term of the time for people troubled by anxiety, fears or compulsions while keeping their grip on reality.

In 1952 Hans Eysenck, a psychologist in London, counted differently. What was missing, he argued, was something to compare with. How many patients recover without treatment?

No proper comparison group existed, so Eysenck used two stand-ins. One was neurotic patients in the state hospitals of New York, who received mainly custodial care: 72 percent were discharged as recovered or improved. The other was five hundred people who had claimed disability insurance for a neurosis. They saw general practitioners, not specialists, and got sedatives and reassurance. After one year 45 percent had recovered, after two years 72 percent, and after five years 90 percent.

Eysenck then gathered nineteen published reports on the results of psychotherapy, covering more than seven thousand cases. Knight's paper was among them. The share recovered or improved was 44 percent for psychoanalysis and 64 percent for eclectic therapy, meaning a mix of methods. The more intensive the therapy, the lower the figure.

Eysenck's conclusion was less sweeping than it is often made out to be. He wrote that the figures "fail to prove" that psychotherapy helps recovery. He did not say they proved it useless. And he listed the reservations himself. About a third of the psychoanalytic patients had broken off treatment, and he had counted them as failures; among those who finished, the figure was about 66 percent. His comparison groups were not true controls. The state hospitals' standard of "improved" was very probably more lenient. What was needed, he said, was properly planned studies.

The yardstick itself was crude. The earlier survey Eysenck relied on admitted that the only unit available was the physician's own report that a patient had recovered or improved. He also quoted other researchers' observation that the more careful a study and the stricter its criteria, the lower the success rate it reports. It is the same problem as the Worcester figures in Part 1, where the recovery rate halved when the superintendent changed.

What had been done, and what had not

First, what had been done. Patients in whom nothing could be found, and in whom doctors had lost interest, now had a form of care: someone listened to them, at length. The idea that a symptom may carry a meaning the patient is unaware of gave mental illnesses that could not be explained from a cause their first coherent story. And the treatment became the mainstream of a whole country's psychiatry.

What had not been done was the checking. The founding case was not cured in the way it was told. The first strong claim about causes was withdrawn by its author. For half a century, recovery rates were reported without a proper comparison with untreated patients. When the comparison was attempted, no difference could be shown.

A little is known about what came next. The properly planned studies Eysenck asked for were carried out in large numbers. In 1977 Mary Lee Smith and Gene Glass pooled nearly four hundred controlled comparisons of treated and untreated people. On average, they reported, the typical therapy client was better off than 75 percent of untreated individuals. But they found no difference in effect between kinds of therapy. That listening and talking help was later confirmed. Whether they help for the reasons Freud gave was not.

Having a satisfying explanation and having a tested one turned out to be different things. The next part turns to those who, in the same decades, came at mental illness from the opposite side, through the body: fever therapy, convulsive therapy and lobotomy.