On the nature and functions
of dreaming
Ernest Hartmann, M.D.
I will outline here my current thinking on the
nature and functions of dreaming based on several pieces of new
research, clinical work with dreams, and a developing view of
the network characteristics of the mind.
I will make some relatively broad statements in outline form followed
by an attempt to back up the statements with available research;
when such research is lacking I will discuss the point using inferences
from research in related areas.
The mind is a net. The cerebral cortex - which,
with some input from underlying areas constitutes the substance
of our minds - functions as a complex net
This is of course a model and cannot be proven by data. There
is always a danger of picturing the mind according to our latest
technological fashions; nonetheless since we think in terms of
visual models I believe we definitely require some image of the
mind on which to hang our thoughts and at this point I believe
the best model involves a net or network of nets. Such network
models have proliferated lately- sometimes optimistically called
“neural nets” in hope that they would correspond with
the actual structure of the cortex and more conservatively called
“connectionist nets” described by computer models assuming
a large number of simple “on/off” units with variable
connections between them. Although these are tremendously over-simplified
nets even a connectionist net of this kind containing only a few
tens or hundreds of units have shown success in modelling certain
human learning tasks - for instance the learning of regular and
irregular verbs (1) or performance on the Stroop Color Naming
Test (2).
I have previously postulated “connecting” and “reconnecting”
as basic aspects of dreaming (3, 4, 5), without specifying any
particular connection mechanism. It appears that the connections
occurring in dreams can fruitfully be examined in such a neural
net or connectionist net model.
Dreaming makes connections more broadly than
waking in the nets of the mind
In this sort of net, all that can happen, during waking, dreaming,
or non-dreaming sleep, is the lighting up of certain patterns
and the strengthening or weakening of the weights on certain connections;
we make connections all the time. I suggest that there is an important
difference: dreaming connects more broadly and more widely than
does waking; in this sense dreaming can be considered “hyperconnective”.
Figure 1 illustrates part of what I have in mind. This is a highly
simplified rendering in two dimensions of a few aspects of the
net using a “spread of excitation” model. I suggest
that in waking there is a tendency for linear development of specific
imagery usually guided by a specific task or goal. For instance,
in thinking of a house, my waking mind seldom pictures a generic
house; rather it is looking for a particular house to answer a
specific question: “Where did I live in 1980?” An entire
pattern lights up representing not just house but very quickly
a specific house in my memory, and in fact, the specific house
in which I lived in 1980. The excitation follows a set pattern;
it remains in a “groove”, with relatively little “spread”.
In dreaming, I suggest the progression is less specific and less
focused. The pattern representing house may be lit up, but then
rather than only moving to a specific house, the excitation process
also spreads “laterally” to patterns representing other
houses and other similar structures — hotels, institutions,
etc. Waking — and for now I am speaking of focused waking
thought, the sort of waking thought that is furthest from dreaming
— tends to stay in a sort of “groove” or “rut”
whereas dreaming thought tends to wander and combine. The setting
for a dream can often be a generic house or a combination of several
houses. In looking over 100 of my recent dreams in which I had
very carefully noted details of the setting, I found that the
most common settings (60%) involved a kind of generic house (or
room or outdoor area); a house that was somewhat like my house,
yet different, a room that was partly a lobby and partly a lecture
hall, etc. Freud’s best known dream likewise starts with
a generic setting: “A great hall...”. These common “generic”
settings would be scored as either “unfamiliar” or “questionable”
settings in Hall and Van de Castle’s standard content analysis
(6); their norms in students are 57% (male) and 53% (female) for
the sum of these two categories. Inge Strauch has recently found
in a group of 11 - 13-year-olds that the setting for their home
dreams as well as lab dreams (REM dreams) were most often unfamiliar
or generic whereas their fantasies usually took place in familiar
settings (unpublished).
This broader and less focused character of dreaming is consistent
with some well-known facts on the biology of the forebrain in
REM sleep, the state during which most though not all dreaming
occurs. I suggested as early as 1973 based on pharmacological
studies that dreaming represents the functioning of the cortex
without the influence of norepinephrine (3). This has been confirmed
and extended: REM sleep appears to be characterized not only by
the virtual cessation of norepinephrine release at the cortex
but by similarly reduced serotonin and increased acetylcholine
(7, 8). The neuromodulatory action of norepinephrine at the cortex
can be summarized as “increasing signal-to-noise ratio”
and “inhibitory sharpening” (9, 10, 11). I view our
hypothetical net as having some more tightly woven or over-learned
portions and other looser portions. The more tightly woven portions
refer to well-learned rapid-processing feed-forward activities
which lead relatively directly from input to output. The less
tightly woven looser regions are those further removed from these
postulated pathways. In Figure 2, I outline, again very roughly,
a model of the brain seen as a growing complexity of interconnections
(interneurons) superimposed on the simplest “reflex arc”
connection between sensory input and motor output. In this simplified
picture I place “feed-forward” mental activities such
as calculating (for instance plotting trajectories in hunting
prey or in catching a baseball) as still relatively close to the
“center”. All our verbal and mathematical abilities,
which activate many different parts of the cortex according to
recent imaging studies, are nonetheless placed relatively “centrally”
in this diagram. The “outer reaches” consist of the
memory nets more readily accessed in reverie, daydreaming, and
dreaming. This is a land populated by moving pictures and by metaphor
(by the potential for producing pictures and metaphor — see
below) with relatively little direct connection to sensory input
or motor output. In this sort of picture, focused waking is more
a hunt and dreaming is more an exploration.
One postulate derived from this view is that certain activities
of focused waking - rapid processing, feed-forward activities,
etc. should be relatively neglected by dreams. This has led to
a study of reading, writing, and arithmetic (the “Three R’s”)
in dreams.
We do not dream of the three R’s
In a preliminary study two scorers simply examined 129 dreams
from two other studies of 68 home dreams and 61 laboratory dreams.
The two blind scorers agreed perfectly that in these dreams there
were zero instances of writing, zero instances of reading, and
one instance of arithmetic described in the dreams (12).
In a broader survey study a questionnaire was sent to 400 good
dream recallers (a mean of 6.8 dreams recalled per month) asking
about any dreams involving the Three R’s (12). In the first
part of the questionnaire respondents were simply asked how often
they had dreamt about reading, writing, typing, and calculating
on a scale going from “never” to “very frequently”.
Respondents were also asked how much time they spent in these
activities during waking. The results were quite consistent for
the four questions; in each case about 90% responded that they
dreamt “never” or “almost never” about the
activity in question, although these subjects reported spending
a mean of six hours per day on these tasks during waking.
Respondents were then asked about the relative prominence of six
activities - walking, writing, talking with friends, reading,
sexual activity, typing - in their waking lives and in dreaming.
The results were very clear-cut (see Table 1). Reading, writing,
and typing were much less prominent than the three other activities
(p <.0001). The activities “writing”, “reading”,
and “typing” did not differ significantly from each
other; nor did the three “other” activities differ from
each other, though they were chosen to cover a wide range of “Non-three
R” activities. All these results demonstrate that we dream
very little of the Three R’s.
The broader connections in dreaming are not
made randomly - the process is guided by emotion
I have outlined above some ways in which dreaming appears to make
connections more broadly than waking, producing “generic”
rather than “specific” imagery - in a more “peripheral”
portion of the nets (further from rapid input-to -output processing
such as the 3 R’s). But is this making of connections a random
process? I think not. In terms of a net such as I have discussed
above, there is a constant flow of excitation and shift of weights.
One can see this as an equilibration, a smoothing out of peaks
and valleys. In an auto-associative net this can be described
mathematically as a settling into a pattern of reduced “computational
energy” or increased “harmony” (13). We can visualize
this roughly as a windswept sea which when the wind dies down
tends to settle towards a relatively smooth surface. This settling
occurs especially when there is less new input and when the net
functions less in a feed-forward and more in an auto-associative
mode — thus in dreaming. The process can be seen as “driven”
by regions of “storm” and high waves — regions
of the net with increased “charge” — or computational
energy. But these waves and wind are not random or meaningless.
I suggest that in everyday human terms they are the emotions and
emotional concerns of the dreamer. The data I have collected (below)
suggests that emotion — the dominant emotion of the dreamer
— is the force which drives or guides the connecting process
and determines which of the countless possible connections are
actualized at a particular time and thus which images appear in
the dream. Dreams “contextualize” the dominant emotion.
Dreams after trauma as the trauma resolves: dreams contextualize
emotion
I have collected series of dreams from over thirty adults and
adolescents who have experienced a severe trauma (14). We are
beginning to analyze these in a quantifiable fashion (see below)
but one qualitative finding leaps out immediately which I call
the tidal wave phenomenon.
People who have experienced any kind of trauma - an attack, a
rape, escape from a fire - in the weeks after the trauma usually
dream to some extent of the actual event, but they also dream
of being overwhelmed by a tidal wave or of being caught in the
path of an onrushing train or being caught by a gang of thugs.
What is happening here? Clearly in these very common dreams the
person is dreaming not about his/her sensory experiences but about
his/her emotional state. The dream seems to be picturing or finding
a context for the emotional state; the dream “contextualizes
emotion”. For instance in the period immediately after trauma,
we have many examples such as the following which seem to contextualize
fear or terror:
• A huge tidal wave is coming at me.
• A house is burning and no one can get out.
• A gang of evil men, Nazis maybe, are chasing me. I can’t
get away.
There are also dreams contextualizing helplessness and vulnerability:
– I dreamt about children, dolls – dolls and babies
all drowning.
– He skinned me and threw me in a heap with my sisters; I
could feel the pain; I could feel everything.
– There was a small hurt animal lying in the road.
For some, guilt is especially prominent and in fact in the longitudinal
series, after dreams that appear to picture fear and terror, there
are dreams dealing with guilt, especially survivor guilt:
– A shell heads for us (just the way it really did) and blows
up, but I can’t tell whether it’s me or my buddy Jack
who is blown up.
– I let my children play by themselves and they get run over
by a car.
– I leave my children in a house somewhere and then I can’t
find them.
In cases where a recent loss - a death of someone close - has
occurred, dreams seem to contextualize grief:
– A mountain has split. A large round hill or mountain has
split in two pieces, and there are arrangements I have to make
to take care of it.
– A huge tree has fallen down.
– I’m in this huge barren empty space. There are ashes
strewn all about.
The above of course have been picked out as examples, simply “illustrating”
rather than demonstrating.
To begin to quantify this difficult area we have recently developed
a scoring system for contextualizing images and have reported
some early results (15). For instance in one group of 135 dreams
consisting of 68 dreams following trauma and 67 dreams without
trauma, in the case of dreams with trauma there were 46 in which
two (or in some cases three) scorers agreed perfectly. Of these
dreams 39 were scored yes (there is a contextualizing image) and
seven were scored no. In the 50 dreams without known trauma where
there was perfect agreement 25 were scored yes and 25 were scored
no (X_ = 13.0; p <.001). The images after trauma were also
scored higher on intensity of the image (2.2 versus 1.8 on a scale
of 1 - 3). In the most “severe” case where eight dreams
were collected in the weeks immediately after a rape, raters agreed
perfectly on scoring eight out of eight dreams. Seven of the eight
were scored yes (definitely having a contextualized image) and
the mean intensity was rated as 2.50. In another sample 42 dreams
as well as 33 daydreams of students with no known trauma were
scored on a blind basis by two raters. The dreams were scored
as containing a contextualizing image in 57% of cases, whereas
daydreams had such an image in only 21% (p <.002). Thus we
are beginning to feel confidence that these contextualizing images
can be scored.
Over a period of weeks or months as the trauma gradually resolves,
the dreams often follow a discernable pattern. First the trauma
is replayed vividly and dramatically but not necessarily in precisely
the way it occurred: there is often at least one major change
in the dream, something that did not actually occur. Very rapidly
the dreams begin to combine and connect this traumatic material
with other material that appears emotionally similar or related.
Often, as we have seen above, a person who has been through one
kind of trauma dreams of all kinds of other traumas that may be
related to this same feeling of helplessness, terror or guilt.
In some cases this connecting involves reactivating previous trauma
and other emotionally important personal themes evoked by the
trauma (“rekindling”). If the dreamer is a “survivor”
while others have been killed or injured, the theme of survivor
guilt almost always emerges. The themes of the dreams and nightmares
are often “Was it him or was it me?” and/or “How
come I survived and he/she didn’t?” (For instance: “...
a shell blows up but I can’t tell whether it’s me or
my buddy who is blown up”; “I get burned in the fire
and my brother’s safe.”). If someone was injured there
is often a theme of guilt in the sense of “did I have something
to do with bringing this on, was I responsible for it?” (This
can occur even if there is absolutely no realistic evidence that
this was the case). The process of connecting the trauma with
other emotionally related material from the dreamer’s life
(and imagination, reading, daydreaming) gradually expands and
takes in more and more other material; the trauma itself plays
a smaller and smaller role and the dreams return to their pre-trauma
state.
The process seems to consist of cross-connecting or interweaving
— making connections with whatever related material is available
in memory and imagination, guided by the dominant emotions of
the dreamer, which gradually become less intense and change their
character as the trauma is resolved or integrated. Although at
the level of the neural net we are still talking of smoothing
out peaks or waves or spreading excitation from over-excited zones,
the process follows definite non-random paths which we can understand
in terms of the dreamer’s emotions and past experiences.
There is often a typical progression of emotional concerns. First,
as noted, the dreams deal with absolute terror and fear; then
sometimes vulnerability and helplessness; then, quite often, the
dreams deal with guilt, especially survivor guilt. Later still
the dreamer deals with grief and loss and attempts to come to
terms with a new reality, etc. These are now individual concerns
of the dreamer and of course they become more complex; I can no
longer give “typical” examples of the imagery used.
Post-trauma dreams are a paradigm - the same
process occurs, though it is harder to discern, in all dreaming
I believe that dreams after trauma, as the trauma resolves (above),
may be a paradigm and an excellent place to start since we know
precisely what emotion must be on the dreamer’s mind. This
is usually less clear in other situations.
I suggest that the same making of connections in pictures and
contextualization of emotional concern can be found in other situations
- probably in all dreams. If one examines dreams in “stressful
situations” in studies such as the classic ones by Breger
et. al. (16) one finds many very similar contextualizing images.
The progression seems to be similar to that after trauma though
without the initial step dealing directly with the trauma.
A huge body of literature on dreams in psychotherapy can be explored
in the same way though of course we must be wary about the selection
process: hardly anyone in therapy (or out of therapy) provides
a complete record of his or her dreams. I have been able to obtain
some relatively complete series - for instance the dreams of a
woman who has recently been overcome by guilt about not being
a good enough mother for her children, reviving a longstanding
childhood guilt of never being able to do anything well enough
to satisfy her demanding parents. I have a record of 35 dreams
told to me by this woman and almost all of them are of the same
kind:
– My children are lost in a storm; I can’t find them.
– I leave my son alone and a big cat is clawing him, killing
him.
– I’m at a hotel by the seashore. My two children are
off in separate rooms and the tide is coming up fast. I wake up
panicked that they’ll drown.
In this very common kind of situation dreams appear to be neither
crazy nor random. They are picturing (contextualizing) a very
clear emotional concern.
Similarly, patients beginning long-term psychotherapy or psychoanalysis
have an obvious concern: “what am I getting into?” or
“what is going to happen to me?”. I am aware of three
different cases in which very similar dreams were reported more
or less as follows:
“I am walking along a mountain path with steep drop-offs
on each side. It is a bit dangerous. There is a large, shadowy
figure accompanying me — I am not quite sure whether this
figure is good or evil.”
Dreams of pregnant women provide another situation in which there
is a series of fairly predictable emotional concerns which can
be traced. Any number of dreams early in pregnancy deal with body
distortions, etc., clearly picturing “what is happening to
my body or will I still be attractive?” Dreams later in pregnancy
deal with small animals or other objects getting larger and sometimes
looking weird and deformed, obviously dealing with the concern
of “what is this baby going to be like”. Finally towards
the end of the pregnancies there are dreams dealing with “will
I be able to be a mother” and “will I be able to take
care of him or her” (17, 18).
I am suggesting that though my recent data come mainly from dreams
after trauma as the trauma resolves, this contextualizing of emotion
may be a more general aspect of the nature of dreams. Dreams contextualize
the emotional concerns of the dreamer. However I suggest that
when one simply looks at a few dreams collected from a few college
students, the dreams may seem confused and almost random since
we know little about the true emotional concerns of the dreamer
(even if we ask for a list) and furthermore the concerns at any
time may be multiple and not very strong.
The form of dreams is metaphor. Dreams can
be thought of as explanatory metaphor
So far, I have suggested that dreams make connections more widely,
more broadly, than waking and that the connections are guided
by emotion and emotional concerns. Dreams contextualize emotion.
But what form do these connections or contextualizations take?
Obviously they do not, or very rarely, take the form of verbal
narratives or mathematical formulas. Though we are often forced
to work with verbal dream reports, we need to keep in mind that
these are only attempts to render the dream experience in a preservable
and reproducible form. What is experienced generally is images
and usually - in sighted persons - visual/spatial images in motion.
The dream-world looks very much like the waking world. The visual/spatial
form of dreams is a fascinating problem in its own right; for
instance, David Foulkes and his associates have studied in detail
how the visual/spatial imagery of dreaming develops gradually
in children at about the same time such imagery develops in waking
life (19).
I do not consider it surprising that dreaming takes the form of
moving visual/spatial imagery because basically that is all there
is. The nets in our minds are made of units and connection weights,
which we cannot directly see or experience, but which represent
the ability to construct, or approximately reconstruct, a visual/spatial
“reality”. This view derives from the basic parameters
of distributed processing: memory is not facsimile but reconstitution.
It is also consistent with work from a totally different direction:
Antonio Damasio (20), based on his clinical neurological work
with brain-damaged patients speaks of knowledge as embodied in
“dispositional representations”. “What ‘dispositional
representations’ hold in their commune of synapses is not
a picture per se, but a means to reconstitute a picture.”
Dreams provide contextualizations in pictures. Another way of
putting this is to say that dreams deal in metaphor. This may
seem far-fetched if one thinks - as many of us were taught - that
metaphor is a specific “trope” used as a rhetorical
device. However, metaphor is ubiquitous. The work of Lakoff, Johnson,
and others (21, 22) has recently demonstrated the ubiquity of
metaphor in our thoughts as well as our language. We can hardly
speak of important topics such as life or relationships without
using metaphor such as a “I am stuck”, “our goals
are in sight”, “we are spinning our wheels”, “it
should be smooth sailing from here on”, “I have to bail
out of this relationship”, etc.
In this sense metaphor is not a rhetorical trope, not even an
aspect of language, but a basic way our thinking is structured.
Admittedly not all our thinking is metaphorical. It becomes increasingly
metaphorical as we move down our continuum from focused thought
to dreaming. When we are calculating for instance or engaging
in a straight-forward over-learned activity such as typing a manuscript
(the 3 R’s), there is little metaphor involved. As we stop
and let out thoughts drift we begin to think in metaphoric pictures;
this occurs even more in dreams. Dreams as we have seen are deficient
at calculating and reading, but they are loaded with (another
metaphor - I can’t help it) pictured metaphor. In fact, this
is not a new idea. Ullman for instance, wrote an entire paper,
called Dreams as Metaphor in Motion, in 1969 (23).
Individual differences: some people function
more “dreamily” than others
There are great individual differences related to dreaming but
these need not concern us here unless they can shed some light
on the connecting and contextualizing aspects of dreaming that
we have discussed above. I believe individual differences relating
to thickness of boundaries may be relevant in this sense. I have
discussed elsewhere in great detail individual differences along
a dimension known as thin versus thick boundaries (24). In general
people with thick boundaries keep everything in their minds separate;
they tend to be solid, well-organized, sometimes rigid; they have
relatively little investment in fantasy; they think in black-and-white
terms. People with thin boundaries have the opposite characteristics:
they tend to merge thoughts and feelings; they have vivid fantasies
not always kept separate from reality; they are less defended;
they think in shades of grey, without black or white.
We have demonstrated that there is a highly significant correlation
between the amount of dream recall and thinness of boundaries;
and in fact persons with thin boundaries not only report more
dreams but their dreams are scored as more vivid, detailed, emotional,
bizarre, “dreamlike”, and with more interaction between
characters compared to dreams of those with thick boundaries (25).
Furthermore, those with thin boundaries have more “dreamlike”
characteristics even in their waking lives. In a preliminary study
a group of students were asked to write down a recent dream as
well as a recent daydream. Reports were scored for bizarreness
and for “dreamlikeness”. Overall dreams were scored
as much more bizarre and more “dreamlike” than daydreams,
but there was a shift between subjects so that those subjects
who scored thin on the Boundary Questionnaire had daydreams that
were just as bizarre and almost as “dreamlike” as the
dreams of those with thick boundaries (26). In other words the
continuum we have discussed running from focused waking mental
functioning at one end to dreaming at the other end is somehow
related to the continuum we have studied across individuals running
from very thick boundaries to very thin boundaries. Those with
thin boundaries live more on the “dreaming” end of the
continuum.
Dreaming may have a function in cross-connecting
or “weaving in” new material - not so much consolidating
material but rather increasing the connections
The above is an outline of my views on the fundamental characteristics
of dreaming and how dreaming differs from waking. But is this
simply the way things are or does dreaming have a function? Starting
again with my collection of dreams after trauma as the trauma
resolves, it sometimes appears that the contextualization of emotion
and making of multiple connections with past similar emotional
material might have a function in terms of connecting or weaving-in
new and difficult material. The initial reaction to a severe trauma
may be something like “HELP! THE WORLD IS ENDING”, “THIS
IS THE MOST HORRIBLE THING THAT HAS EVER HAPPENED”, “HOW
CAN ANYONE SURVIVE THIS”, but then as connections are made
there may be a gradual increase of “YES, THIS FEELS BAD BUT
IT’S A BIT LIKE....”, “I’VE EXPERIENCED SOMETHING
LIKE THIS, I’VE WORKED ON THESE FEELINGS”, or “I’VE
DEALT WITH SOMETHING SIMILAR; IT’S HARD BUT IT’S POSSIBLE
TO CONTINUE”.
In this sense dreaming may have a quasi-therapeutic function:
the making of connections in a safe place. I have reviewed many
similarities between dreaming (whether or not remembered) and
the process of psychotherapy, especially after trauma (27). Both
good psychotherapy after trauma and dreaming first provide a safe
place for work to be done. In therapy the safe place is much more
than the physical setting; it involves the safe “boundaries”
of the therapeutic situation and the gradual trusting alliance
formed between patient and therapist. In dreaming - especially
in REM sleep - the safe place is provided by the well-established
muscular inhibition which prevents activity and the acting out
of dreams. Once a safe place is established the therapist allows
the patient, especially the traumatized patient, to go back and
tell her or his story in many different ways, making connections
between the trauma and other parts of the patient’s life
- overall making connections and trying to integrate the trauma.
As connections are made between the terrible recent event and
other material, the emotion becomes less powerful and overwhelming
and the trauma is gradually integrated into the rest of life.
Dreaming may perform at least some of these same functions. Dreaming
and psychotherapy both can be conceptualized as “making connections
in a safe place”. Thus dreaming may have a quasi-therapeutic
adaptive function which can be seen most easily after trauma though
I believe again that trauma is a paradigm and that dreaming has
the same function, though less easily discernible, at other times.
In terms of the nets of the mind, the spreading out of excitation
or reduction of “computational energy” is useful in
presumably allowing the net to function better, in a more harmonious
state. But the effect is not purely “energetic”; the
spread of excitation forms increased connections and cross-connections
which inevitably alter the future functioning of the net. The
trauma, or any disturbance, is cross-connected, “woven in”
by dreaming as numerous new connections and contexts are provided.
This process is likely to be useful for future functioning since
a new trauma or disturbance will be less serious, will produce
less “storm-waves” since appropriate contexts and cross-connections
are already present.
In this broad sense I see dreaming - making of broad connections
and contextualizing - as having a function which can be seen both
as restorative/adaptive in an immediate sense (spreading excitation,
calming the storm) and as producing changes in memory networks
which are adaptive for the future. This change in networks is
not a consolidation of memory but a broadening of memory through
cross-connections - an increase in connections, a weaving in of
new experience.
This possible function of dreaming can be called quasi-therapeutic
or adaptive and I would call it a contemporary theory of the function
of dreaming rather than my own theory since others have suggested
very similar functions starting from very different data bases.
Thus French and Fromm (28) and Palombo (29) using clinical data,
Breger, et. al. (16) studying acutely stressful situations, and
Cartwright (30) in research on people under a prolonged stress
(divorce), have all proposed versions of an adaptive function
of dreaming. Jones (31) made an analogy between dreaming and effective
psychotherapy. Koulack (32) has proposed a complex functional
theory which at least in part involves “mastery of stress”.
Fiss (33) has proposed that dreams function to maintain “self
structures”. Greenberg and Pearlman (34, 35) have suggested
several versions of an adaptive problem-solving function. Milton
Kramer (36) has proposed a “selective mood regulation”
function of dreaming which derives from very different studies
but is similar to what I have been discussing in terms of a calming
of stormy seas or spreading out of excitation.
Although dreaming should not be confused with REM sleep, since
most dreams come from REM sleep any hypothesis on the nature and
functions of dreaming should at the very least be compatible with
what we know of the nature and functions of REM sleep. As we saw
above, the present view of dreaming is very compatible with low
norepinephrine availability at the cortex during REM sleep providing
less “inhibitory sharpening”. Concerning function, although
the functions of REM sleep are still unknown, the present views
of the functions of dreaming in terms of cross-connections is
certainly compatible with two related views on the functions of
REM sleep. It fits well with the view by Roffwarg, et. al. (37)
that REM sleep, especially in young organisms, helps to “develop
the nervous system” - evidently by making new connections.
It is also compatible with the view that REM sleep functions in
the “repair, reorganization, and formation of new connections
in amine-dependent forebrain systems” summarized as “knitting
up the raveled sleave of care” (3).
Critics skeptical about any function of dreaming often question
how dreams can be important if most of them are forgotten. I suggest
that remembering or not remembering an individual dream image
is not what is important in terms of function. What is probably
most important is the making of broad cross-connections in the
net, the redistribution of weights, etc., all of which can occur
whether or not the actual dream content is remembered. Of course,
when a dream is remembered, then in addition to this basic function,
dreams can be useful in any number of other ways - in problem
solving, self- knowledge with or without a therapist, and occasionally
in helping with artistic and scientific discovery. And although
I do not believe the essential function of dreaming requires recall,
yet it must of course be admitted that all our conjectures about
dreaming - including the present one - are necessarily based on
examination of the subset of dreams that are remembered.
References
1. Rumelhart, D.E., & McClelland, J. L. (1986). On learning
the past tenses of verbs. In J. L. McClelland, D. E. Rumelhart,
and the PDP Research Group (Eds.), Parallel Distributed Processing,
Volume 2:Psychological and Biological Models. Cambridge, MA: MIT
Press.
2. Cohen, J. D., Dunbar, K., & McClelland, J. L. (1990). On
the control of automatic processes: a parallel distributed processing
account of the stroop effect. Psychological Review, 97, 332-361.
3. Hartmann, E. (1973). The Functions of Sleep. New Haven, CT:
Yale University Press.
4. Hartmann, E. (1976). Discussion of ‘The Changing Use of
Dreams in Psychoanalytic Practice’. The International Journal
of Psychoanalysis, 57, 331-334.
5. Hartmann, E. (1991). Dreams that work or dreams that poison:
What does dreaming do? Dreaming, 1, 23-25.
6. Hall, C. S., & Van de Castle, R. L. (1966). The Content
Analysis of Dreams. New York: Meredith.
7. Hobson, J. A. (1988). The Dreaming Brain. New York: Basic Books.
8. Siegel, J. M. (1993). Brainstem mechanisms generating REM sleep.
In M. H. Kryger, T. Roth, W. C. Dement (Eds.), Principles and
Practice of Sleep Medicine. Philadelphia, PA: W. B. Saunders Company,
125-144.
9. Foote, S. L., Freedman, R., & Oliver, A. P. (1975). Effects
of putative neurotransmitters
on neuronal activity in monkey auditory cortex. Brain Res., 86,
229-242.
10. Woodward, D. J., Moises, H. C., Waterhouse, B. D., Hoffer,
B. J., & Freedman, R. (1979). Modulatory action of norepinephrine
in the central nervous system. Federation Proceedings, 38, 2109-2116.
11. Servan-Schreiber, D., Printz, H. W., & Cohen, J. D. (1990).
A network model of catecholamine effects: gain, signal-to-noise
ratio and behavior. Science, 249, 89- 895.
12. Hartmann, E. (1996). We do not dream of the Three R’s:
a study and implications. Sleep Research, 25, 136.
13. Smolensky, P. (1986). Information processing in dynamical
systems: foundations of harmony theory. In D. E. Rumelhart, J.
L. McClelland, and the PDP Research Group (Eds.), Parallel Distributed
Processing, Volume 1. Cambridge, MA: MIT Press.
14. Hartmann, E. (1996). Outline for a theory on the nature and
functions of dreaming. Dreaming, 6, 147-170.
15. Hartmann, E., Rosen, R., Gazells, N., & Moulton, H. (1997).
Contextualizing images in dreams - images that picture or provide
a context for an emotion. Sleep Research, in press.
16. Breger, L., Hunter, I., & Lane, R. W. (1971). The Effect
of Stress on Dreams. New York: International Universities Press.
17. Garfield, P. (1988). Women’s Bodies, Women’s Dreams.
New York: Ballantine Books.
18. Maybruck, P. (1989). Pregnancy and Dreams. Los Angeles, CA:
Jeremy P. Tarcher.
19. Foulkes, D. (1971). Longitudinal studies of dreams in children.
Science and Psychoanalysis, 19, 48-71.
20. Damasio, A. R. (1994). Descartes’ Error. New York: G.
P. Putnam’s Sons.
21. Lakoff, G. (1993). The contemporary theory of metaphor. In
A. Ortony (Ed.), Metaphor and Thought. Cambridge, England: Cambridge
University Press.
22. Lakoff, G., & Johnson, M. (1980). Metaphors We Live By.
Chicago: University of Chicago Press.
23. Ullman, M. (1969). Dreaming as metaphor in motion. Archives
of General Psychiatry, 21, 696-703.
24. Hartmann, E. (1991). Boundaries in the Mind: A New Psychology
of Personality. New York: Basic Books.
25. Hartmann, E., Elkin, R., & Garg, M. (1991). Personality
and dreaming: the dreams of people with very thick or very thin
boundaries. Dreaming, 1, 311-324.
26. Kunzendorf, R. G., Hartmann, E., Cohen, R., & Cutler,
J. (to be published). Bizarreness of the dreams and daydreams
reported by individuals with thick or thin boundaries.
27. Hartmann, E. (1995). Making Connections in a Safe Place: Is
Dreaming Psychotherapy? Dreaming, 5, 213-228.
28. French, T. & Fromm, E. (1964). Dream Interpretation. New
York: Basic Books.
29. Palombo, S. (1978). The adaptive function of dreams. Psychoanalysis
and Contemporary Thought, 1, 443-476.
30. Cartwright, R. (1991). Dreams that work: the relation of dream
incorporation to adaptation to stressful events. Dreaming, 1,
3-10.
31. Jones, R. M. (1970). The New Psychology of Dreaming. New York:
Grune and Stratton.
32. Koulack, D. (1991). To Catch a Dream. Albany, NY: State University
of New York Press.
33. Fiss, H. (1986). An empirical foundation for a self psychology
of dreaming. Journal of Mind and Behavior, 7, 161-192.
34. Greenberg, R., & Pearlman, C. (1975). A psychoanalytic
dream continuum: the source and function of dreams. The International
Review of Psychoanalysis, 2, 441-448.
35. Greenberg, R., & Pearlman, C. (1993). An integrated approach
to dream theory: contributions from sleep research and clinical
practice. In A. Moffitt, M. Kramer, R. Hoffmann (Eds.), The Functions
of Dreaming. Albany, NY: State University of New York Press.
36. Kramer, M. (1993). The selective mood regulatory function
of dreaming: an update and revision. In A. Moffit, M. Kramer,
R. Hoffman (Eds.), The Functions of Dreaming. New York: State
University of New York Press.
37. Roffwarg, H. P., Muzio, J., & Dement, W. C. (1966). Ontogenetic
development of the human sleep-dream cycle. Science, 152, 604-619.
Ernest Hartmann, M.D.
Tufts University School of Medicine; Sleep Disorders Center, Newton-Wellesley
Hospital
Ernest Hartmann, M.D.
27 Clark Street
Newton, MA 02159
Phone (617) 965-5872
Fax (617) 965-6548
>>Up
