Having Nightmares

To the Editors:

In his review of my book, The Nightmare (Basic Books, 1985), Dr. Charles Rycroft [NYR, March 28] tries to blur the carefully documented distinction I make between nightmares and night terrors, suggesting that the latter may be the tail ends of nightmares “too terrifying to be remembered in detail.” However, there is neither experimental nor clinical evidence for this view, and considerable evidence against it. For instance, if it were true, one would expect a night terror sufferer to go through a period of nightmares as his condition improves; this is not what happens.

Rather than take up other specific points about nightmares, I would like to discuss an issue of broader interest raised by Dr. Rycroft: The methods to be used in advancing our knowledge of the mind. He twice uses the unusual word “faith” to describe my reliance on sources of data other than his preferred source—the patient in psychoanalysis or psychotherapy (“…his faith in the reliability of psychological tests and surveys…” and “one would have to have greater faith than I do in psychological tests and psychiatric interviews to take seriously formulations based on such shaky foundations”).

I plead innocent. I have not placed excessive “faith” in any one source of data, be it tests, interviews, or psychoanalytic case material, though I have made use of all of these. I hold some ordinary skeptical scientific beliefs: the data examined and the methodology used depends on the question being asked; controlled studies are desirable whenever possible; and conclusions are strengthened if similar results are obtained from independent sources.

To illustrate, let me outline in simplified form the questions I asked and the methodologies I employed to answer them in arriving at my conclusions about nightmares. First, in my therapeutic work with a few patients and in another project, it seemed to me that there was something unusual about people who reported frequent nightmares; they were singularly open and vulnerable; they did not employ the usual defense mechanisms. Here I used the methodology of listening to patients on the couch and in the consulting room to get the beginnings of an idea. Would Dr. Rycroft have me stop there? Should I write a treatise on nightmares based on these few cases? Or should I wait—years perhaps—for other patients with frequent nightmares to arrive at my door? These patients are rare, and psychoanalytic research is limited: We can only follow where the patient leads—we cannot introduce new material or ask questions to further our own interests.

I chose a different path. Having formulated the question—“What are the personality characteristics of persons with frequent nightmares?”—I tried to answer it systematically. This involved detailed psychiatric interviews by myself and others with a large number of persons—not patients—who had frequent nightmares. On the basis of these interviews, I confirmed and expanded my notion that these were unusual people: They were not neurotic; they had a vulnerability to schizophrenic illness and they were artistic and creative people. I was often told that these were interesting clinical findings, but wasn’t there some way of “pinning them down,” or quantifying these impressions. This required another methodology: I made use of the long, old-fashioned, but eminently quantifiable questionnaire known as the Minnesota Multiphasic Personality Inventory (MMPI). Using this instrument, I showed that people with frequent nightmares indeed had unusually high scores on the scale relating to schizophrenia (and on several other so-called “psychotic” scales), whereas they did not have high scores on the scales usually associated with neurosis.

One possible criticism was that perhaps these unusual traits were characteristics of research subjects, of “newspaper an answerers,” rather than of persons with nightmares. To answer this objection, I performed another study on similar subjects but with two control groups of subjects who also answered newspaper ads: one group with no nightmares but very vivid dreams and another group who reported neither nightmares nor vivid dreams. By comparing these groups, I was able to show that the MMPI results and the interview results characterized people with nightmares but not the other two groups of newspaper and answerers.

Based on my detailed interviews, I developed the general concept of “thin boundaries”: The people with nightmares had unusually “thin boundaries” in many different senses. This led to another question and still another methodology: I predicted that the “thin boundaries” would show up on psychological tests in which one can get some semi-quantitative measure of “boundaries.” The best known such test is the Rorschach (ink blot) test. In my second study, each person in the three groups mentioned took a Rorschach test, scored by a psychologist blind as to group membership. The tests were scored for a number of standard Rorschach scores but also for “thin boundaries”—responses which included amorphousness (“a spreading ameoba-like shape”), merging (“Two women merged together”), thinness (“a thin nightgown”), or penetration (“a frog being stabbed”). The nightmare group had significantly higher scores than the control groups on this boundary scale, while there was no difference on the other Rorschach scales. Here again, a different methodology confirmed impressions obtained from clinical situations and from interviews.

These studies led to further questions which required still other methods, including studies of a large group of veterans with nightmares, a survey of schizophrenic patients, and a large survey of students to examine the relationship between nightmares and artistic interests. Certain other questions required sleep laboratory methodology, including a study of chemical factors that made nightmares more frequent. And in my chapter on the interpretation of nightmares I relied on patients with nightmares whom I have seen in psychotherapy and psychoanalysis, since this appeared to be the only appropriate source of data.

In other words, I employed, and described in detail, a number of methods and data sources. I tried to use the most appropriate methodology for answering each question and whenever possible to confirm results obtained in one way by results obtained in another. I am myself a psychiatrist and a psychoanalyst. I have respect for the analytic situation both in therapeutic terms and as a source of data; but there are other sources. In my book on nightmares, I let the reader know explicitly what data were used to approach what questions rather than asking him to have faith in conclusions based on my “vast clinical experience.” Our knowledge of the mind is scientific knowledge. Marvelous and complex is its conception, gestation, and development. Not all wisdom is born on the couch!

Ernest Hartmann, M.D.

Tufts University School of Medicine

Boston, Massachusetts