Summer 1998

<< table of contents  

T H E   D I S T I L L E R   O F   D R E A M S

With a novel statistical system,
a psychologist opens up a new avenue
to the meaning of dreams.

By Michael Hagmann

A physiology of dreams?

IN 1953 neurobiology took command of dream research. The discovery of REM sleep - a certain phase of sleep named for the dreamer's rapid eye movements - marked the beginning of a new era in the study of dreams (and sleep). The emphasis on physiology gave rise to new ideas about dreams, the most radical of which consider dreams to be mere trash cans for our sleeping brains.
      For Nobel laureate Francis Crick, of the Salk Institute in San Diego, dreams are nothing more than random attempts to clear from the brain unneeded or even harmful memory. Crick claims this clearing activity is a necessary step to reset the brain for the next day, much as one erases old data from a floppy disc before reusing it.
     Not quite brain garbage, yet still randomized products created by uncontrolled nerve-cell fireworks in our sleeping brain - that is how J. Allan Hobson, a Harvard psychiatrist and neurophysiologist, describes dreams. In his view, dreams come into being when these random signals reach the cortex, the brain's center of thinking and reasoning, which tries hard to make sense of the arbitrary inputs. Random firings, Hobson says, are due to a very different composition of signal-transmitting chemicals, called neurotransmitters, in the brain during REM sleep as compared to the waking brain. According to Hobson, a different brain chemistry also explains - through chaotic nerve impulses - our wild and bizarre dreamscapes during REM sleep.
     However, the great majority of the dreams Domhoff analyzed contain familiar settings and known characters that take part in recognizable, and reasonable, activities. "Dreams are, in fact, generally well-organized and lack bizarre elements," agrees David Foulkes, a professor Emeritus of psychology at Emory University in Atlanta.
     Nor can dreams simply be equated with REM sleep, which is what both Crick and Hobson focus on, Domhoff says. The equation between REM sleep and dreams was the prevailing credo among dream/sleep researchers in the late 1950s and early 1960s. For the first time, psychologists could trace a mental process - dreams - to a physiological activity - REM sleep. "It was a very exciting time and an exploding field. We were the hot guys on the block - even newspapers wanted to talk to us," Domhoff remembers. But the simple equation did not stand the test of time. The longer psychologists studied sleeping subjects, the more it became obvious that dreaming occurs in all other phases of sleep, Domhoff says.
     Domhoff and Foulkes are particularly aggravated by neuroscientific attempts to explain the nature of dreams by reducing them to some physiological phenomena, such as the measurement of brain activities. Says Foulkes, "No PET scan, or any other fancy biotechnical doodad, is capable of observing or characterizing the nature of dreaming." They refer to such approaches as "biologic reductionism."
     "I am always very wary about neurobiological findings," Domhoff says. "Dreams are not just eye movements. There might be a neurobiology of dreaming, but never a neurobiology of dreams."

THINK OF DREAMS and you might think of Sigmund Freud and Carl Jung, of sex and phallic symbols. Of Freud's "royal road to the unconscious" paved with weird metaphors and obscure symbolism. Maybe of psychedelic hallucinations, the visual counterpart of Pink Floyd's "Dark Side of The Moon." Or even of ties to another world. Some might remember Gregory Peck being haunted by surrealistic nightmares in Alfred Hitchkock's "Spellbound," with its ingenious production by Salvatore Dal’, the master of burning giraffes and melting watches. Whatever dreams are, most of us view them as enigmatic visions from the abyss of our mind.
      When psychologist Bill Domhoff thinks of dreams, he sees them as short tales or theater plays with coherent story lines, no more bizarre or obscure than our routine waking thoughts. But most of all, he thinks of an Excel 5 spreadsheet brimming with numbers and a bar diagram he calls h-profile. That is Domhoff's unspectacular way of trying to quantify and thus decipher our nocturnal movie shows.
      What Domhoff finds in thousands of dreams dreamt all over the world is a surprising regularity in dream content that underlies individual and also cross-cultural differences. Besides, his results show a striking consistency in what we dream of throughout our adult lives. In Domhoff's view, it is through dreams that our sleeping mind expresses our daily concerns and ideas about ourselves and the reality that surrounds us.

DOMHOFF, who collects dreams at the University of California, Santa Cruz, shares with Freud and Jung the firm belief that dreams offer the most direct path to the dreamer's personality. But other than their conviction that dreams have meaning, Domhoff's notions differ radically from those of the founders of psychoanalysis. For both Freud and Jung, dreams were puzzling accumulations of strange scenarios flooded with arcane props, as if the cameraman or the director of the movie went berserk.
      To Domhoff, most dreams do not appear bizarre or alien at all. "I used to think of dreams as much more mysterious and symbolic because I shared many of the cultural beliefs back in the Sixties, when Freud was all-pervasive. But the more we looked at dreams, the more regular and the more consistent they got," Domhoff says. "In a way [our research] demystifies Freud and Jung."
      And dreams need not be fantastic or absurd to tell us something about the dreamer, Domhoff contends. Buying groceries, not finding what you want and losing your car keys are just as telling as dream scenarios. But what does dreaming of Safeway reveal about our inner selves?

TO DECODE a dream, Domhoff begins - oddly enough - by coding it. At the core of this process figures a tabular system. It features a variety of categories, such as characters (say, people or animals), objects (perhaps a car), settings, social interactions between them (an act of aggression for instance) and various other classes and subdivisions, such as misfortunes, emotions and striving. Using this system which was developed by pioneering dream researchers Calvin Hall and Robert Van de Castle in the 1960s, Domhoff deconstructs the content of a dream or preferably of a collection of dreams, a so-called dream series. For example, the dream fragment, "My parents sat in the audience when I sang a Cole Porter song," would translate into: 1MFA/1FMA-2JUA-1MPA. Then, Domhoff simply counts each single occurrence of an image, or dream element. What he ends up with is a sheet full of numbers, an alpha-numeric translation of a dream.
      "Dreams are stories that we tell ourselves at night. Our coding system is a way of turning a text into a set of numbers or, in other words, turning symbolic behavior into scientific data," Domhoff says. This systematic and quantitative approach makes dreams comparable across the world, Domhoff says.
      Since the late 1940s, Hall, Domhoff and others have used this system to analyze thousands of dreams of American college students. Viewed collectively, they form the so-called "norm," the average dreamer's dream or frame of reference, with which other dreams can be compared. College students, being quite abundant at a university, became the first dream Guinea pigs for Hall and Domhoff - a simple matter of convenience (the researchers could just as well have chosen firefighters or opera singers). A norm dream is characterized by how often friends, strangers, cats, dogs, screwdrivers, mud slides, sexual escapades, broken legs, etc. occur. Individual dream elements are then assigned to one of the various categories. Thus, an average female dream consists of 5.3 objects, 1.3 settings, 2.8 characters, 0.6 friendly and 0.7 aggressive interactions (all with further subdivisions, of course). Findings in the norm group have been replicated quite accurately four times with dream series gathered independently from different dreamers within a period of forty years. Hence, there seem to be only minor changes, if any, in what comparable groups of people dream over the years, Domhoff says.

THE THEORY behind the Hall/Van de Castle system and its predictions about the dreamer assumes continuity between dreams and our waking thoughts. Both are manifestations, so the theory goes, of our mind in action, an idea that dates back to ancient Greece. Aristotle was the first to state "dreaming is thinking while asleep." Hall and Domhoff couldn't agree more. For Hall, the images of our dreams are "the embodiment of thoughts." And Domhoff says all the results Hall, himself and others have obtained so far, seem to support this so-called cognitive theory of dreams, first proposed by Hall in the early 1950s. "It's about us and our thinking of the world while we sleep," Domhoff says. "Dreams are the most completely subjective picture of what's on our mind. They are extremely revealing about our self-conception."
     In this cognitive dream theory, the frequency with which dream elements appear is thought to reflect the concerns and the interests of the dreamer. Thus, any obvious deviation from the norm, or any other group of reference, should allow predictions about the dreamer's preoccupations, concerns, conflicts, and his or her conception of the surrounding world while he or she is awake. Once the dreams are analyzed, Domhoff scrutinizes his inferences, and hence his dream theory, in an interview with the dreamer.
      And since the Hall/Van de Castle system is, at large, statistical in nature, the more dreams Domhoff has, the more accurate his predictions. "I don't think I can tell you much about a single, isolated dream. But about 70 dreams [of the same person] - that's a psychological X-ray or a fingerprint. With this in hand, I am confident that I can tell you a lot about that person," Domhoff says.
      Dream research ˆ la Domhoff has come a long way since Freud's Vienna days. Domhoff's coding headquarters is a white, office-like room with a neutral, almost clinical atmosphere - not very dreamy (as some of us might say). No couch, not even a dreamer in sight. Just the reports scattered all over. Most of them are computer printouts looking like manuscripts - which is what they are, dream manuscripts, screenplays of our sleep. One of them is a photocopy of a dream diary, enriched with hand-drawn sketches of what the dreamer encountered on his mental excursions. A detailed railroad engine dominates one of the pages. On another one a naturalistic drawing of a beetle figures alongside the dream report.
      The stacks of dream reports are interleaved by yellow, handwritten sheets. These sheets contain the rules by which Domhoff and his colleagues code the dream reports. Cryptic columns of letters and numbers on the left are facing short explanations, reminiscent of a handwritten vocabulary for a foreign language. (Quite a peculiar language, one might say. A8, for example, stands for "an aggressive act resulting in the death of a character," or 1ANI would mean my dog.) Everything you could possibly dream of is listed here. It is a bare-bones guide to dream content analysis.
      Following these rules, Domhoff takes apart dream series of individuals, as well as of certain groups like skydivers, vegetarians, or gay men. He compares them to the norm, but also to each other or to any other group, for differences and similarities in the various categories. Adam Schneider, one of Domhoff's colleagues, recently developed a computer program for these sophisticated statistical analyses. The result is a bar diagram that mathematicians arbitrarily term the h-profile. It portrays the deviations of the analyzed dream(s) from any given reference group. Each bar corresponds to one of the various dream elements. With one glance, Domhoff sees how dreams differ between individuals or groups.

ONE OF THE early results obtained with this system was what Domhoff refers to as "astonishing regularities in dreams from all over the world". The general types of characters, interactions and elements in dreams are much the same worldwide. One common thread in our dreams is, for example, that aggression prevails over friendliness and misfortune over good fortune. But clear-cut gender-specific, developmental, cross-cultural, and individual differences also emerge.
      For example, men are more likely to dream about other men (almost 70 percent of all characters), whereas women are less discriminatory (both genders figure around 50 percent). Also, men show more aggression in their dreams than women do. So, the h-profile of male dreams shows large bars (in the positive direction) on the Male/Female percent and the aggression scale relative to female dreams.
      Dreams also differ in several ways between children and adults. Animals play a more prominent role in children's dreams (up to 50 percent of all characters, compared to only 5 percent in adults' dreams). So do outdoor activities. Besides, children are more often the victims of aggressive behavior rather than the source. After adolescence, however, dreams don't change significantly with age. When Domhoff compared dreams within three individual dream series that extended over 20 to 50 years, he found surprisingly little change in an individual's dream history.
      Perhaps the most telling result using the Hall/Van de Castle system concerns cross-cultural differences in dreams. "[Individuals from] tribal hunter-and-gatherer societies dream much more about animals [than the US college student norm], and that's exactly what you would expect. In Japanese dreams, animals are really low, which reflects the life in big cities where you're not allowed to keep animals," Domhoff says. And while he admits that these are not very profound insights, "it tells us we're on the right track" using their quantitative system of content analysis. "I think of dreams as the 4 C's: concern, conception [of ourselves and the surrounding reality], continuity [between our waking thoughts and our dreams], and consistency [throughout our adult lives]," Domhoff says.
      Domhoff has learned, however, that dreams do not always reflect our daily behavior or actions. Compelling examples are dream series with a high overall incidence of murders or highly sexual dreams. (The latter, in contrast to common belief, make up only about 10 percent of all reported dreams, including dreams with hugs or kisses.) When Domhoff analyzed these dream series he, of course, predicted that the dreamers were aggressive and sexually active, respectively. In interviews with two of them afterward, this turned out not to be the case. However, both dreamers admitted that, in their thoughts and day-dream fantasies, they would often experience these feelings. At the moment, Domhoff says he still has trouble separating dreams that actually reflect real behavior from the ones that "only" reflect thoughts and ideas.
      According to Domhoff, this example also shows the validity of the Hall/Van de Castle system, which completely relies on what Domhoff calls "blind analysis." "The less we know about the person the better," Domhoff says. "There are a lot of conclusions we draw when we only see or hear somebody." To find meaning in dreams, Domhoff looks only at the dream reports themselves and analyzes the content statistically without any interpretation. He does not rely on - or even want - free associations, biographical or any other information provided by the dreamer. The interview with the dreamer is Domhoff's measure of all things. "When we ask them, people often realize their fears and their concerns. We are usually right [with our inferences], and even when we're wrong we learn something [about our theory]," Domhoff says.

EVEN MORE challenging for Domhoff are dream series where the Hall/Van de Castle system fails completely, where it leaves an apparently irreconcilable contradiction between Domhoff's predictions and the dreamer's self-perception or self-portrait. In separate interviews with friends or family members of the dreamer, Domhoff tries to bridge the gaps. "If some of the psychological traits are confirmed in outside interviews we might have discovered a blind spot of the dreamer, a misconception of him- or herself for example, an unrecognized rivalry with an older brother. A "conflicting" content analysis somehow is a sign for self-deception - that is, the dreamer is not aware of a certain attitude, feeling or emotion," Domhoff says. But even though dreams might contain unconscious material, he says "the process [of dreaming] itself is first and foremost a cognitive process during sleep." This is as close as Domhoff gets to the realms of psychoanalysis or the unconscious and its "inventor," Freud.
      In contrast to Domhoff, both Freud and Jung concentrated mainly on a specific subset of dreams to develop their dream theories. Freud was fascinated by sexual dreams, which he considered wishes in disguise. These fantasies are usually repressed by our consciousness when we are awake. Jung focused on universal mythical symbols, and he came to believe that dreams reveal undeveloped parts of our personalities. For Jung, we dream what we do NOT think about.
     Extended dream series, without any preselection or bias, do not support either of these theories, Domhoff says. But even Domhoff admits that there are some dreams - often the most captivating, puzzling and memorable ones - in which we create highly metaphorical pictures. In collaboration with linguists, Domhoff tries to nail down and analyze these metaphorical elements in dreams, such as "exposure is embarrassment," "up is feeling good," "hot is anger," "bridge is transition," or a "fallen" woman. His goal is to see whether their general meaning in our colloquial language and their appearance in dreams fit his theory that dreams express how the dreamer conceives the reality he or she is living in.
     "Ultimately, it's these very metaphorical dreams that we want to understand," Domhoff says. "The danger lies in doing the same thing as Freud and Jung, who were mere interpreters of isolated metaphors and symbols." A metaphoric analysis has to be embedded in the context of a person's 'regular' dreams, Domhoff thinks. Information obtained with the Hall/Van de Castle system about a certain dreamer, he says, should provide effective insurance against Freudian over-interpretation, which resulted from Freud's belief in the unconscious as the creative force behind our dreams.

NOT EVERYONE in today's psychological community has chucked the unconscious altogether. Veronica Tonay, a clinical psychologist also at UC Santa Cruz, studies the connection between dreaming and personality traits, especially creativity. She still sees plenty of room for the unconscious.
     To a large extent, her findings support Hall's theory about the continuity between dreams and waking thoughts and concerns. This is not as much a surprise for Tonay now as it was when she first began her analyses. "I started out thinking that Hall's theory was too reductionist. I wanted to prove him wrong," she said. But after years of experience with the Hall/Van de Castle system she changed her mind. "It's absolutely amazing how much you know about a person after coding their dreams," Tonay says.
     The one level of our personalities that she says does not fit the continuity theory concerns the emotions we experience while dreaming. She observed, for example, that sad people had predominantly cheerful dreams and vice versa. "Feelings in waking life and in dreams are not necessarily the same. Often this is a subconscious process," Tonay says. She agrees with Jung in that we dream about the emotional side within ourselves which is unknown to us or repressed during our waking life.

TONAY notwithstanding, many psychologists have left the Freudian/Jungian track these days. A major turning point was the discovery of a certain phase of sleep called REM sleep (see Sidebar), during which most dreaming was thought to occur. But after an initial euphoria, psychologists slowly discovered that dreaming occurs throughout our sleep. Their hopes of finally understanding dreams on a physiological level were dashed. For Domhoff, a promising period ended in disillusion. "From 1965 it went straight downhill. Many of us felt we had lost our reason for being in the lab. A whole world started collapsing around me," Domhoff says.
     So, Domhoff changed gears completely in the late 1960s and made a 20-year-long leap into sociology, studying the political and social power structure in the United States. It was only toward the end of the 1980s that he turned back to his earlier love, the field of dreams.
     Having collected several thousand dreams, Domhoff is as firmly convinced that there is meaning in dreams as he is that there is no psychological function to them. "Most people who claim a function for dreams, for example in the consolidation of memory, in fact talk about REM sleep. My best guess is that there is no function in dreams because it seems we don't really need them," Domhoff says, referring to brain injury patients who lost the ability to dream without suffering any other obvious aftermath. Even Allan Hobson agrees on this point. "I think we need to have REM, but I don't think we need to dream," he says.
     We may not need dreams, but most people would not want to lose them either. Whatever their function or meaning will turn out to be, they will always fascinate us. After all, there is a (bit of a) dreamer in everyone. Through our dreams, life with all its adventures and (mis)fortunes thrills us even - and in some cases especially - while we sleep. Life just keeps going on behind closed eyes, it seems. As the late Havelock Ellis, the English psychologist and writer, put it: "Dreams are real while they last. Can we say more of life?"