Two Boston scientists, one from the laboratory bench and one from the bedside,
are tackling an age-old question: Why do we dream and what does it mean?
by Lisa Birk
A man dreams his mother is riding a tiny tricycle back and forth across a
"Love me, love me, love me," she sings operatically.
psychoanalyst, the dream might spotlight two aspects of the mother, emotional
neediness and a penchant for drama.
To a neurophysiologist, the dream indicates a decrease in norepinephrine and a
decrease in serotonin, leading to an increase in acetylcholine, which produced
random images in the man's mind. Just like every other dream.
How we think about our dreams has always reflected how we think about our
place in the universe. In ancient Greece, dreams were messages from the gods.
In biblical times, dreams were prophecies. In contemporary America, experts are
at odds -- are dreams a meaningful system? or are they a random process, a
kind of brain tic? -- and have been for almost five decades.
The split is drawn vividly here in Boston, where psychiatrists Ernest Hartmann
and J. Allan Hobson, former colleagues at a Harvard teaching hospital,
have been rivals in the field of dream theory for nearly 40 years. Now they are
battling it out again, if indirectly, on the bookshelf. Hartmann, a
psychoanalyst, Tufts med-school professor, and director of the sleep-disorders
center at Newton-Wellesley Hospital, describes his approach in his 1998 book,
Dreams and Nightmares: The New Theory on the Origin and Meaning of
Dreams (Plenum). Allan Hobson, a Harvard med-school professor and director
of Harvard's neurophysiology lab, describes his latest approach and its
implications in his book Consciousness, published in November by
Both men are challenging Freud's dream theory, which is to say they are
challenging not only the notion of the unconscious, but also talk therapy, free
will, the muse, the nature of creativity, and the nature of madness. They are
challenging, in other words, our reigning image of ourselves. But they are
challenging that image in very different ways, and no one is yet sure what will
be erected in its place.
The way we think about dreams today was shaped nearly 100 years ago by a
neurologist named Sigmund Freud, who decided neurology was too primitive to say
much of anything about dreaming. Freud invented modern psychiatry with his
paradigm-breaking book The Interpretation of Dreams, which, among other
things, proposed the idea of the conscious and the unconscious mind.
For nearly 50 years, psychiatry students were trained in Freud's theory of the
id, ego, and superego. They were trained in Freud's theory of the unconscious,
Freud's theory of free association, Freud's theory of therapy, and Freud's
theory of dreaming. Dreams were central to this system; Freud himself had
called dreams "the royal road to the unconscious," meaning that dreams were the
best path to the otherwise hidden and unknowable recesses of the mind. During
sleep, he believed, our rational mind (the ego) relaxes, allowing instincts
(the id) to surface. However, if our forbidden wishes or drives became too
explicit, we'd wake up. Thus, he hypothesized, dreams were coded: they occurred
in symbols, and the psychoanalyst's job was to decode the dream, and so permit
the dreamer insight into his true self.
Then, in 1953, came a seismic shock. Two researchers at the University of
Chicago discovered that sleepers experienced bursts of rapid eye movement in
their sleep. The scientists also noted that dreaming happened consistently and
perhaps exclusively during these REM bursts, which suggested that the
actual catalyst of dreams was not repressed memories or desires but, rather, a
physiological process. Even animals, it seemed, had REM sleep.
Suddenly dreams, which according to Freud were an upwelling of powerful
suppressed urges and anxieties, were being described as "automatic." Dream
content might not be terribly important or even terribly individual -- because
dreams were triggered not by one's horrible past, but by the release of a
If the scientists who discovered REM sleep were right, if dreams were merely a
chemical process, then the unconscious was -- to put it bluntly -- a bunch of
hooey. Talk therapy was a bunch of hooey. Among psychiatrists, the more
scientifically inclined retreated to one shore and studied the brain. The
psychoanalytically inclined retreated to the opposite shore and studied the
mind. Scholarly articles took one side or the other. The world of
psychoanalysis, which had tenuously bridged two worlds -- the medical and the
psychological -- began to pull apart.
It was at this moment that Ernest Hartmann and
J. Allan Hobson entered Harvard Medical School to finish their training as
The year Allan Hobson turned eight, he and his father built a lab for Allan's
experiments. The family lived in West Hartford, Connecticut, near Avon
Mountain, and Hobson frequently went foraging in the woods.
Sometimes he brought back snakes. Hobson was afraid of snakes, partly because
he knew the woods were full of them (some of them poisonous), and partly
because local wisdom had it that even if you bashed a snake's head in or
chopped its head off, it wouldn't die until after sundown. Not only could
snakes kill you; it seemed they also had supernatural powers.
Hobson investigated. His first set of experiments consisted of catching snakes
and bashing their heads in. He'd take each snake back to the lab and wait for
its body to stop wriggling. Sundown would come, and then his bedtime would
come, and then sometime in the night the snake stopped twitching.
Hobson conducted a second set of experiments, which varied little from the
first except that Hobson would wake himself periodically through the night to
check on the snake to see exactly when it had stopped moving. He then conducted
a third set of experiments, which again resembled the first except that Hobson
began chopping the snake bodies into smaller and smaller bits. Still, each bit
wriggled for hours.
Eventually Hobson figured out -- whether through further experiments or by
reading books he no longer remembers -- that the snake's nervous system is
segmented. The body twitches not because snakes are supernatural but because
their physiology is set up that way.
By high school, Hobson was routinely dissecting brains, a hobby many of his
classmates considered "weird, perverted and disgusting," he writes in the
acknowledgments of his 1988 book, The Dreaming Brain. But adults,
especially scientists, encouraged him. The first to do so was a dyslexia
specialist named Page Sharp, who took an interest in the 15-year-old Hobson as
a subject. Tests showed that Hobson had been born with dyslexic brain wiring
but functioned normally. Sharp wanted to know why. Eventually, Sharp was so
impressed with Hobson's intelligence that he hired him as a research
Some years later, Jack Ewalt, Harvard's charismatic chief of psychiatry at
midcentury, must have seen something similar: an energetic, smart young
scientist who was willing, even eager, to buck the trends. Ewalt partially
funded Hobson's lab at Harvard's prestigious Massachusetts Mental Health Center
when Hobson was still a lowly post-doc.
Hobson made good use of the funds and the lab. In 1973, 20 years after REM
sleep was first linked with dreaming, he shook the psychoanalytic world with a
lecture titled "The Brain as Dream Machine." No doubt the word "machine" was
intended to rankle those who still believed that dreaming was a most human act
-- that dreaming distinguished us from animals, and definitely
distinguished us from machines.
Much of Hobson's research was based on cat brains, and many critics thought
his presentation premature. Four years later, Hobson answered the critics with
a flourish (some called it grandstanding): he opened "Dreamstage," a six-week
"multi-media portrait of the sleeping brain" at Harvard's Carpenter Center.
Crowds thronged in to see a human sleeper sleep, to see the machine that
recorded the sleeper's brain waves. (The brain-wave patterns, it turned out,
looked a lot like a cat's brain-wave patterns.) To some scholars, Hobson's
exhibit bordered on the tacky. To the public, it was sexy and fun -- and who'd
ever thought brain science could be fun? Hobson was a force to be reckoned
But Hobson was not the only student singled out by Jack Ewalt in the early
'60s. Ewalt also believed in another bright, promising young man, a Viennese
named Ernest Hartmann.
Like Hobson, Hartmann was attracted to science. He remembers with distaste
a psychiatry professor who urged students not to read books but just to
"sit with the patients." Unlike Hobson, however, his approach to analysis was
not shaped by chopping up snakes. It was shaped by, well, analysis.
Ernest Hartmann's father, the renowned Viennese psychoanalyst Heinz Hartmann,
was an early and loyal disciple of Sigmund Freud. The junior Dr. Hartmann
(at age two) even met Freud, a fact he includes on the jacket flap of his
Hartmann remembers a poll taken in the '50s or '60s, in which both
psychoanalysts and the public were asked to name the best psychoanalyst in the
country. The public voted for a doctor named Erik Erikson. The psychoanalysts,
however, voted for Heinz Hartmann.
Hartmann senior was the psychoanalysts' psychoanalyst, and so loyal a Freudian
that people nicknamed him the "son of Freud." So Ernie Hartmann is, he
sometimes jokes, the grandson of Freud.
Meeting Hartmann is a bit like meeting Freud's grandson, or your mental
image of Freud's grandson. He doesn't have the beard or the paunch, but his
wispy hair sometimes stands up or out, the way really smart people's hair is
supposed to. He has a Viennese accent, a courtly, almost old-fashioned
kindliness, heavy red drapes, and a Newton home office lined floor-to-ceiling,
wall-to-wall with books. In the middle of the office, he's placed two red
leather chairs on either side of an elegant table stacked with thick books and
a big rock incised with one word: DREAM.
Hartmann, however, is not such a loyal grandson when it comes to ideas. In
decades of research, both as a scientist and as a psychoanalyst, he has come to
reject nearly half of Freud's tenets. Hartmann does not believe, as Freud did,
that dream content is irrational or even coded. He rejects the idea that all
dreams are based on a repressed wish. What distinguishes Hartmann from most
followers of Freud, according to J. Christian Gillin, a professor of
psychiatry at the University of California at San Diego, is that Hartmann "is a
more independent thinker." Hartmann has not just modernized Freud's theory, he
has rebuilt it.
Hartmann goes along with neurophysiologists who say that dreaming is
chemically activated; he's done some experiments along those lines himself. He
just doesn't think that's very interesting. He wants to understand why we dream
and what those dreams mean, and what the answers imply for a new model of the
Hartmann spent years as a researcher and a psychoanalyst working with
traumatized people, and over time he collected 40 dream series containing from
six to a thousand dreams each. The dreams were collected in the weeks or months
after a trauma or, in some cases, over five years.
The nightmares of traumatized people, he found, had certain common features:
there was a strong feeling, often expressed in metaphor, and there were bits of
the dreamer's past woven in.
Traumatized people often dreamed not so much about the facts of what had
happened as about the way the event made them feel. This was particularly true
as time went on. Sometimes people dreamed of the trauma itself -- say, a rape
-- but the details, the setting, or the characters would be different. Later,
people would dream about other traumas they hadn't experienced -- tidal
waves, for instance -- but the dream rape and the dream tidal wave would
engender a common feeling of terror.
Big losses triggered similarly metaphoric dreams. Hartmann
writes of a man whose powerful mother had just died. In his dream, "This huge
mountain has split apart and there are pieces lying around. I am supposed to
make arrangements to take care of it." In the dream of a woman who had just
lost her mother, "There was an empty house, empty and barren. All the doors and
windows were open and the wind was blowing through."
These dreams, Hartmann theorized, were driven by emotion, but the content was
metaphoric, and the metaphors were important. The metaphors, he writes, were a
way of showing the dreamer, "It is a terrible event, yes, but is it unique?
Let's look at . . . other catastrophes. . . . Awful,
yes, but it's part of a whole catalog of human disasters, . . . [and]
sometimes people survive." Dreams, in other words, are a way of healing the
Many of the nightmares he studied drew on previous experience. Hartmann cites
an example of a World War II vet who had a recurrent nightmare of a
severed human head lying on the side of the road. He'd awaken horrified and
angry. The dream puzzled the vet: for all the atrocities he'd witnessed, he
couldn't recall having seen a severed human head. Then, after some therapy, he
remembered an event from childhood that had conjured the same feelings. His pet
guinea pig had disappeared, and some time later he found the guinea pig's head
in the evening's stew. His mother was trying to supplement their poor diet.
Hartmann believes this sort of dream, the kind that draws on past experience,
is an example of the brain saying, "Let's look at other times you felt
terrified. Not quite the same? No, but let's keep looking; wasn't there some
similar feeling? And you survived."
Dreams, in this conception, may be the mind's equivalent of fibroblasts. After
a cut, fibroblasts knit together the edges of the skin. If trauma similarly
rips the mind, then dreams, by making connections to other traumas and to past
experience, help knit the psyche back together.
Hartmann collected another type of dream, too: those of scientists and
artists. Among them is the story of inventor Elias Howe, who for five years had
tried to invent a sewing machine. Howe could visualize most of the machine, a
needle piercing fabric over and over, but he couldn't figure out how to get the
machine to hold and knot the thread. Then one night he dreamed he was a
missionary caught by natives carrying spears. The spears, he noticed, had holes
in the tips. When he woke, he had the solution to his problem: a hole through
the tip of the needle to hold and knot the thread.
Robert Louis Stevenson claimed to have dreamed the entire plot of The
Strange Case of Dr. Jekyll and Mr. Hyde. More than
50 percent of mathematicians answering a survey said they had solved a
problem in their dreams at least once. Hartmann lists famous artists, ranging
from Albrecht Dürer to William Blake to Frida Kahlo, who recorded a dream
as the source of their inspiration.
Hartmann began to think about the common properties of these two sorts of
dreams: nightmares that made metaphors of trauma, and dreams that found
solutions to problems. He began to think of dreaming as "hyperconnective," a
state of mind in which we are able to make more associative leaps than we can
in the waking state. And then, appropriately, he dreamed a visual
representation of such a connective sort of brain. He dreamed of "a pile of
Persian carpets loosely tied together."
As he developed the theory, Hartmann began to visualize the mind as composed
of "neural nets" lying one on top of another. Imagine a pile of tiny
electrified fishing nets. The spot where one rope knots with another is a
neural connection. In the conscious mind, dreams are more likely to stay within
a net; in dreams, they are more likely to leap from net to net. The mind,
Hartmann believes, makes connections all the time, whether awake, asleep, or
dreaming -- but the dreaming mind connects "more broadly, more widely."
"The waking mind is on a hunt," says Hartmann. It is goal-directed. "The
dreaming mind is on an exploration." The dreaming mind's spreading connections
help people think associatively, creatively, as Howe did with the sewing
machine, and Stevenson did with Jekyll and Hyde.
Allan Hobson will have none of it. "My challenge to the metaphor merchants," he
says, rolling close in his office chair, "is, prove it! Show us scientifically!
The whole point of science is to check belief. Rhetoric? I'm not interested.
Literature? I'm not interested. Science is insurance against being fooled."
Hobson rolls around in his chair a lot, often aggressively, to emphasize a
point or to ask a question. It is easy to imagine his delight nearly 30 years
ago over that lecture title, "The Brain as Dream Machine," the title that
tweaked establishment psychoanalysts and launched his career.
Hobson has graduated from a home lab to the neurophysiology department of
Harvard Medical School, but in some sense he is still, at 60-plus, the boy who
cut up snakes to challenge local wisdom. He spends his nights experimenting: he
sometimes wears a "nightcap," a device made of a modified tennis headband with
sensors for eye and body movement, to record his own REM sleep.
Then again, the man is not quite the boy. When he was around 40 he began
keeping a dream log because, he says, he realized that he was going to die.
Pressed on the subject of why a dream researcher, after nearly two decades of
investigating dreams, finally starts recording his own, he has a variety of
answers, ranging from "it's fun" to "it's an existential benefit." In 20-some
years, he's recorded five or six thousand dreams in 102 volumes. And his
published books are full of accounts of his own dreams.
Nonetheless, as a scientist he's not persuaded that he is recording messages
from his unconscious. Hobson thinks that the science necessary to prove
Hartmann and other psychoanalysts right is missing. In some cases, such as his
"dream splicing" experiment, he believes the evidence flat-out contradicts
their dream theories.
Dream splicing was an experiment designed by Hobson and Robert Stickgold, an
assistant professor of psychiatry at Harvard Medical School, in which
assistants took a variety of written dream reports, cut them up where the
scenes shifted, and spliced them together at random. They left an equal number
of dream reports whole. They then asked a panel of judges, all psychiatrists,
to determine which dream reports had been spliced and which had been left
His theory is that if dreams were truly individual, psychiatrists should be
able to identify dreaming styles and thereby piece dream reports together
accurately. The psychiatrists on the panel couldn't do it, and neither could
Hobson. He concluded that dreams don't tell a coherent or useful story:
that the metaphoric content dreamers find in their dreams, and the insight some
dreamers gain upon waking, is false.
"Psychoanalysts' dream theory," says Hobson, "is the most fanciful, delicious,
absurd theory that anyone's taken seriously for 100 years." And they've taken
it seriously, Hobson believes, because it's so seductive. So comforting. How
thrilling, after all, to believe that your subconscious creates while you
sleep, that your subconscious heals psychic wounds, that something wiser than
your conscious self is in charge.
In place of Freud's theory and Hartmann's theory and all the other
psychoanalysts' theories, Hobson offers his own: a purely physiological theory
that he calls "activation-synthesis."
In its simplest form, activation-synthesis, originally formulated by Hobson
and his colleague Robert McCarley, another psychiatry professor at Harvard,
comes down to this: we dream because of two processes, one called activation
and the other called synthesis.
In the "activation" phase, the brain stem -- the lower, more primitive part of
the brain -- blocks most incoming sensory data (smell, touch, sound) and also
blocks motor impulses (the impulse to run, for instance). This is why someone
dreaming about monsters does not physically react by running away. A dreamer is
The cycles of REM sleep are governed by the balance of three
neurotransmitters: serotonin, norepinephrine, and acetylcholine.
(Neurotransmitters are chemicals that communicate messages from nerve cell to
nerve cell.) While the body sleeps, the brain stem's production of serotonin
and norepinephrine decreases almost to the point of inactivity. This raises the
level of acetylcholine, which induces REM sleep -- and excites the visual,
emotional, and motor centers that may be responsible for flooding the sleeping
brain with images. Certainly, Hobson believes, there is a correlation between
increasing acetylcholine and dreaming, but he can't yet say definitively that
the one causes the other.
"Synthesis" takes place in the cortex, the seat of thinking and the
higher part of the brain. The brain responds to the incoming flood of images by
knitting them together as best it can, turning the random bits of data into a
story, or what we know as a "dream." The reason dreams seem coherent to the
dreamer -- at least at the time -- is because serotonin and norepinephrine, the
chemicals that shut down to trigger the dreaming process, are also the
chemicals that govern judgment. In other words, Hobson and McCarley believe,
the very mechanism that enables dreaming prohibits us -- at least while we are
dreaming -- from realizing how arbitrary our dreams really are.
In some sense this sounds anti-humanist. The process is automatic, the content
random. Still, Hobson sees dreaming as a possible agent of creativity. We are
nightly visited with spectacular visions -- "literary productions," as he puts
it -- which, if we are good editors, we can transform into something else: a
dream log, maybe even art.
Besides, Freud's theory of dreaming, says Hobson, is boring.
Psychoanalysis is so limited. Everything is self-referential. "There are more
things," he says, "than are dreamt of in your philosophy, Mr. Freud."
But the frontiers of brain science are rapidly expanding, and even as Hobson
distances himself from Freud's theories, others are distancing themselves from
Hobson's. Cognitive scientist David Foulkes, in a 1996 article for the journal
Sleep, reviewed the last 40-plus years of dream research and concluded
that dreaming happens not just in REM sleep, but also in other phases.
That idea may challenge some of Hobson's most fundamental principles. "What
I'll conclude for you," says Gerald Vogel, a critic of Hobson and director of
the sleep-research laboratory at Emory University, "is that REM sleep is not a
necessary condition for dreaming, and it is not a sufficient
condition for dreaming. So therefore, unique physiological events can not
explain dreaming. That's it in a nutshell, kid."
And if REM's "unique physiological events" cannot explain dreaming, then
certainly, says Alfred Margulies, an associate professor of psychiatry at
Harvard Medical School, "it's untenable to say that [neurophysiology] proves
Freud wrong. Dream physiology is a lot less clear than folks would have it."
Still, most experts believe Hobson has made a significant contribution to our
understanding of REM sleep. J. Christian Gillin believes that Hobson and
other neurophysiologists have also strengthened traditional psychoanalysis. "In
the '40s and '50s," says Gillin, "there were virtually no clinical trials ever
conducted to determine the effectiveness of psychoanalytic theory. Today,
nearly every theory is being tested empirically." So it seems that each side is
at long last building a new methodological bridge, one that spans the conundrum
of the brain-mind.
However, the two camps are still divided over content. Neurophysiologists
believe that understanding the brain is our best shot at explaining the mind.
Psychoanalysts believe the neurophysiologists are focused on the least
interesting part of the story.
As an illustration, Ernest Hartmann claps loudly, startling a visitor.
"There," he says, "I activated your cortex."
What's important at such times, he insists, is not that your brain was
activated, but what you were thinking. What images came into your mind?
To Hartmann, those questions are the key to understanding the cortex, the
higher-level brain, and, ultimately, human nature. And that's where Hartmann
thinks pure neurophysiology is lacking.
"All we can say now [about dreaming from a biological perspective]," says
Hartmann, "is that [the brain] lights up differently during waking and
dreaming. We can say the amygdala (the center of emotion) lights up during
dreaming, but. . . . " He shrugs as if to say, So what?
Margulies agrees. Studying neurophysiology to the exclusion of psychology "is
like a deaf Martian studying . . . how the human auditory system
works. But the Martian could never understand Mozart."
Gillin has another metaphor: If dreaming is a drama, then
"[neurophysiologists] have discovered where the house manager sits, and who
turns on the lights, but they don't tell you what play will go on tonight, or
who the actors will be."
"Psychology is a higher level of biology," says Hartmann. "Maybe someday
biology will be able to describe what's happening in the cortex. But it will
take 10,000 statements to say `defense mechanism.' Psychology is the shorthand
way, and at the moment the best way, to describe what's going on in the
In the end, many researchers in both fields believe that the overriding goal of
dream theorists is not to pick sides, but to find one unifying theory that
explains both brain and mind. That job may very well take professionals working
from both the lab and the couch.
For all their differences, Hobson and Hartmann exemplify the trend toward
unification. Each expresses a kind of intellectual joy upon hearing the other's
theory of creativity. When Hartmann learns that Hobson believes dreaming allows
for a random recombination of data that "opens up the system for creativity,"
the psychoanalyst sits back, looks skyward, and says, "I like that."
When Hobson is told of Hartmann's theory of dreaming as a "hyperconnective"
state, he pauses, quieted, as if rolling the concept around his mind. "It's a
genial theory," he says.
The argument between them is a central question in our age. Biology or
metaphor? Science or art? Which discipline better explains the "truth" of
dreaming? The debate being conducted here in Boston, in our heads and often
over them, is a debate not only over dreaming, but over how we should be
thinking about the big questions.
"I suspect in the long run," says Gillin, "we'll have to reconcile both
theories." But in the meantime, the two camps are still a long way apart.
At the very end of one long interview, Hobson told me thoughtfully, maybe even
wistfully: "All complex systems are noisy. Why shouldn't the brain be noisy?"
He meant that dreams might be the static in a complex system, not the music.
Hobson is still going into the woods, confronting the local wisdom, chopping up
the snakes to figure out what makes them wiggle and what makes them stop.
Hartmann, in one of his last calls to me, said, "You left your scarf here,"
and then he paused. "You know what Freud would say? Freud would say you didn't
want to leave." And metaphorically speaking, he was right.
Lisa Birk is a freelance writer living in Cambridge.