Freud is widely considered to be one of the "giants" of the modern human sciences. Is this a justified evaluation - did he really provide new and important insights into the human mind? Or is he rather more of a charlatan, better at self-aggrandizement than self-analysis?

A recent book edited by Frederick Crews argues convincingly for the latter conclusion, making the case not only that Freud was a fraud, but also that the entire structure of modern psychoanalysis is little more than a house of cards. In his introduction, Crews asserts early on that psychoanalysis has almost no factual support and was developed on little more than Freud's say-so.

An excellent example of this is the famous case of Anna O., a woman who was allegedly cured of hysteria in the 1880s by Freud's mentor Josef Breuer when he used hypnosis to retrieve her repressed traumatic memories. Freud reported that her symptoms had "immediately and permanently disappeared," but he was lying. Both he and Breuer knew that she had only gotten worse under Breuer's care. Over the course of six years following her treatment, Anna was committed at least three times - once by Breuer himself.

But isn't it true that at least a few features of modern psychoanalysis have a ring of truth? Certainly - various defense mechanisms like projection certainly sound like they could be true, and our daily interactions with others seem to provide anecdotal support. However, under Freud's theories of these mechanisms, there is no way to tell when an utterance or action is "literal," and when it is a sign of something else entirely.

This is not the only instance of this flaw in his ideas - rather, it is representative of a common thread throughout his writings where there are no clear, objective standards by which psychoanalytic assertions can be independently tested and evaluated. The advantage for analysts is that this prevents them from ever encountering "uninterpretable" material. Yet this advantage is purchased at steep price.

The first cost is internal schisms and strife: because any utterance or behavior can in theory be interpreted in a multitude of ways without obvious error, it is easy for analysts to disagree and form their own "methods." The second cost is psychoanalysis' status as a science. Some have recognized this and have tried to pass psychoanalysis off as something else entirely; but Freud himself insisted that his theories were science, and there is no way that he was correct.

Of course, Freud's confidence in himself and his ideas were seemingly boundless - no new data could shake his certainty that he had discovered something important. It is true, of course, that his theories "explain" various observations, yet that isn't enough. A good scientific theory can't simply "explain" things, but must instead also explain why it is superior to competing explanations:

To distinguish between them one must be able to isolate that one circumstance which will make a difference. But this is precisely what is impossible if one bars the use of experimental manipulation in principle. A reliance on data acquired by observation rather than experimental control is a reliance on data determined by circumstances as they happen to occur rather than as they are required by the demands of a good scientific test.

From an article by Frank Cioffi, we learn a set of characteristics which distinguish a pseudoscience from a genuine science:

...first, a liberality in adding ambiguities and escape clauses to the doctrine's claims, so that no seeming disconfirmation can ever be decisive; second, a weakening of the very idea that self-contradiction is a sign of trouble; third, an appeal to the proviso that only practitioners of the doctrine's special method of knowledge acquisition are entitled to evaluate its results; fourth, recourse to quasi-physical terminology that creates an impression of precision while actually allowing "the quantitative factor" to serve as an excuse for the failure of predicted effects to materialize; and finally, the maintenance of a pretense, despite all evidence to the contrary, that independent validation has already been demonstrated.

Cioffi then proceeds to demonstrate, along with the other authors, that Freudian theory fits those characteristics very well.

Criticizing Freud is not something that will help you win a popularity contest. His followers are not too far off from whom Eric Hoffer would characterize as "True Believers" - they are committed to the Freudian program do not accept dissent from his key theories.

Throughout the course of the book, the contributors effectively demonstrate the validity of three arguments. First, there is no empirical basis to prefer Freud's psychoanalysis over various competing theories regarding the mind and human behavior. Second, the method of gathering evidence in psychoanalysis is circular, failing to provide independent confirmation of the claims being made. Third, psychoanalysis makes no predictions which can be independently tested and evaluated. These lead us to the inevitable conclusion that psychoanalytic theory is a pseudoscience which should simply be dispensed with.