The Scientific Study of Dreams 2003

TheScientific Study of
DREAMS
J'lcunil Networks, Cognitive Development,
and Content Analysis
G. W illiam Domhofj
The Scientific Study of
DREAMS
N euralNetworks, Cognitive Development,
and Content Analysis
G. W illia m D om hoff
A M E R I C A N P S Y C H O L O G I C A L A S S O C I A T I O N
W A S H I N G T O N , DC
C opyright © 2003 by the A m erican Psychological A ssociation. A ll rights reserved.
Except as perm itted under the U n ite d States C opyright A c t o f 1976, no p a rt o f this
p u b lica tio n may be reproduced or distributed in any form or by any means, or stored in a
database or re trie va l system, w ith o u t the p rio r w ritte n perm ission o f the publisher.
F irst P rin tin g , N ovem ber 2002

Second P rin tin g , July 2003
Published by
A m erican Psychological Association
750 F irst Street, N E
W ashington, D C 20002
www.apa.org
T o order In the U .K ., Europe, A fric a , and the M iddle

A P A O rder D epartm ent East, copies may be ordered from
P.O. Box 92984 A m erican Psychological Association
W ashington, D C 20090-2984 3 H e n rie tta Street
T e l: (800) 374-2721 C ovent G arden, London
D ire ct: (202) 336-5510 W C 2E 8LU England
Fax: (202) 336-5502
T D D /T T Y : (202) 336-6123
O n lin e: www.apa.org/books/
Em ail: order@apa.org
Typeset in Goudy by W o rld C om position Services, In c., S terlin g, V A
P rinter: P ort C ity Press, Baltim ore, M D

C over Designer: Berg Design, A lbany, N Y
T echnical/P roduction E ditor: K risten R. S u lliva n
The opinions and statements published are the responsibility o f the authors, and such
opinions and statements do n o t necessarily represent the policies o f the A m erican
Psychological Association.
L ib ra ry o f Congress C atalog ing -in -P ub lica tion D ata

D om hoff, G . W illia m
The scie n tific study o f dreams : neural networks, cogn itive developm ent, and co n te n t
analysis / by G. W illia m D om hoff.— 1st ed.
p. cm.
Includes bibliographical references and index.
ISBN 1-55798-935-4 (alk. paper)
1. Dreams. 2. Dreams— Physiological aspects. 3. Dream in terpretatio n. I. T itle .
BF1091 .D66 2002
154.6'3— dc21
2002026114
B ritis h L ib ra ry C atalog uin g-in-P u blicatio n D ata

A C IP record is available from the B ritish Library.
Printed in the U nited States o f A m erica

W ith thanks to the greatest Dream Team o f them all:
Adam Schneider, Sarah D unn, Melissa Bowen, H eidi Block,
Thomas Van Rompay, and Ryan H arvey.
CONTENTS
Acknowledgments ................. ................ .....................................
Introduction: The N eurocognitive Approach to Dreams .... ............... 3
Chapter 1. Toward a N eurocognitive M odel o f Dreams .. ............... 9
Chapter 2. M ethodological Issues in the Study

o f Dream C ontent ..................................... ............. 39
Chapter 3. The H a ll-V a n de Castle System .... ............... ............. 67
Chapter 4. A New Resource fo r C ontent Analysis .......... ............. 95
Chapter 5. New Ways to Study M eaning in Dreams ....... ........... 107
Chapter 6. A C ritique o f T raditional Dream Theories .... ........... 135
References .... ......................................... ............. .................. ...... ........... 171
Index ........... ........................................... . .. ......... ............ ........... 197

A b o ut the A u th o r .................. .............. .............................. ........... 209
vii
ACKNOWLEDGMENTS
Thanks to David Foulkes fo r two careful readings o f the entire manu

script; to Adam Schneider fo r his superlative w ork in creating the tables
and figures in the book; to Teenie M atlock and Raymond Gibbs, Jr., fo r
th e ir help in developing the ideas about figurative th in kin g in dreams; to
Sarah Dunn, H eidi Block, and Melissa Bowen fo r th e ir many helpful analyses
o f the dream series th a t is discussed in chapter 5; to Thomas Van Rompay
fo r the analyses o f the dream series discussed in chapter 4; and to Ryan
Harvey fo r his contributions to the analysis o f the Barb Sanders dream series
in chapter 5.
A special thanks to Robert S tickgold and Richard Zweigenhaft fo r
th e ir many useful comments on parts o f chapters 1 and 6; to Linda J.
Mealey and Nicholas S. Thompson fo r comments and suggestions th a t
greatly improved the discussion o f the accuracy o f dream reports in chapter
2; to K elly Buckley for his suggestions on chapters 5 and 6; and to Raymond
Gibbs, Jr., fo r his reassurances on the discussion o f cognitive theory in
chapter 1. I also w ant to thank M ark Solms fo r answering many questions
about the neural netw ork fo r dreaming and Lawrence W ich lin ski fo r back
ground inform ation on the role o f neurochemicals in dreaming.
DREAMS
INTRODUCTION:
THE NEUROCOGNITIVE
APPROACH TO DREAMS
B uilding on established descriptive findings in three separate areas o f

dream research, this book attempts to show th a t it is now possible to take
the study o f dreaming and dreams to a higher level th a t can rig h tly be called
scientific. For reasons th a t become apparent throughout the book, it has
n o t been easy fo r inquiries in to dreams to escape from anecdotal clin ic a l
theories, on the one side, and neurophysiological reductionism , on the other.
I t is only since the 1990s th a t new neuropsychological and neuroimaging
findings can be combined w ith replications concerning the lack o f dreaming
in young children and the consistency o f adult dream content to suggest
the new neurocognitive synthesis th a t is presented here.
In addition to proposing a neurocognitive model o f dreaming, the book
provides a m ethodological foundation fo r testing some aspects o f the model
and includes new findings on dream content th a t demonstrate the po ssib ilit
ies o f those methods. The model is an open-ended and tentative one; it is
meant as a starting p o in t fo r future research. The work upon w hich it builds
has come to m aturity and a possible integration at an ideal moment, because
advances in neuroscience, content analysis, cognitive linguistics, statistics,
and computer software make it possible to test the new model.
The tim ing may be rig h t for consideration o f a new model because
the em pirical findings from systematic studies do not support any o f the
best know n clin ic a l theories. This p o in t has been made in several reviews o f
the accumulating research evidence (D om hoff, 1999a; Fisher & Greenberg,
1996; Foulkes, 1996a, 1999). However, the fu ll case against a ll o f these
theories is brought together fo r the first tim e in chapter 6 o f this book. This
critique is placed at the end o f the book so th a t much o f the em pirical
evidence on w hich is it based can be presented first.
This approach also differs from the neuropsychological model called
activation-synthesis theory, w hich starts at the level o f single-cell firings
in the brain stems o f lesioned cats and moves to speculations about how
3
the brain stem may have direct effects on dream content (Hobson, 1988;
Hobson & M cCarley, 1977; Hobson, Pace-Schott, & Stickgold, 2000b). By
way o f contrast, the neurocognitive model suggested in this book begins at
the phenom enological level w ith the subjective experience o f dreaming and
w ith waking dream reports, then proceeds to a neuropsychological level.
The sim ilarities and differences between the two models are explored
throughout the book and in chapter 6.
The model presented in this book starts w ith findings from neuropsy
chological assessments o f people w ith brain lesions who report changes in
th e ir patterns o f dreaming (Solms, 1997). These neuropsychological findings
are corroborated and expanded by neuroimaging studies o f rapid eye move
m ent (REM ) sleep, the stage o f sleep during w hich dreaming most v iv id ly
occurs (Braun et ak, 1997, 1998; Heiss, Pawlik, Herholz, W agner, & W ien-
hard, 1985; M aquet et al., 1996; Nofzinger, M in tu n , W iseman, Kupfer, &
M oore, 1997). This w ork allows the model to explain the origins o f dreaming
and to glimpse the outlines o f the neural netw ork th a t underlies this form
o f cognition.
N ext, the model adds findings on children who were observed over
many hundreds o f nights in the sleep laboratory (Foulkes, 1982, 1999;
Foulkes, H o llifie ld , S ullivan, Bradley, & Terry, 1990). This w ork allows the
model to explain how dreaming develops in terms o f the cognitive processes
th a t make it possible. This w ork combines w ith the findings on the neural
netw ork to make the model tru ly neurocognitive.
Finally, the model explains the nature o f dream content by incorporat
ing insights drawn from quantitative investigations o f many thousands o f
dream reports using a rigorous system o f content analysis (D om hoff, 1996;
H a ll & Van de Castle, 1966). This coding system rests at the nom inal level
o f measurement, uses percentages and rates to correct fo r differences in the
length o f dream reports, and has high re lia b ility . It is one o f the few coding
systems th a t has been used extensively by investigators other than those
who created the system, including researchers from Japan, India, Switzerland,
and the Netherlands, and it has proven useful w ith dreams collected by
anthropologists in small tra d ition a l societies as w ell (D om hoff, 1996).
The results from these three em pirical areas o f dream research make
it possible to state the follow ing generalizations that lead to a neurocognitive
model and many testable hypotheses:
■ Dreaming occurs when a relatively specific neural netw ork—

located p rim a rily in the lim bic, paralim bic, and associational
areas o f the forebrain— is activated in the absence o f external
input coupled w ith a le ttin g go o f control by the self. I f defects
occur in this netw ork, dreaming can be lost tem porarily or
4 THE SCIENTIFIC STUDY OF DREAMS

permanently, or it can be impaired in some way, such as through
the loss o f visual dream imagery.
■ Dreaming is a cognitive achievement that develops gradually
over the first 8 or 9 years o f life.
■ The “ output” o f the conceptual systems in the neural netw ork
for dream generation, called dream content and available to
scientists through w ritte n or transcribed dream reports, is drawn
from many o f the same schemata and memory systems as waking
thought and is generally continuous w ith waking conceptions;
this content includes a great deal o f previously unrealized repeti
tio n in characters, social interactions, misfortunes, negative
emotions, and themes.
These generalizations lead to several hypotheses concerning the links

between the neuropsychological, developmental, and content analysis
findings:
■ Because specific neural defects can lead to the loss or im pairm ent
o f dreaming and because dreaming develops gradually in c h il
dren, adultlike dreaming may depend on the m aturation o f the
neural netw ork for dreaming. For example, the complete loss
o f dreaming in adults due to injuries in or near either parieto
tem poral-occipital ju n ctio n , along w ith the finding th a t in
creased dream reporting in young children correlates w ith visuo-
spatial skills, together suggest th a t the a b ility to dream in c h il
dren may depend in part upon the development o f the neural
netw ork fo r spatial construction, w hich is centered in this
region.
H The importance o f lim b ic structures in dreaming and the repeti
tive nature o f dream content suggest that the “ em otional brain”
mapped out by LeDoux (1996) may be one basis fo r the repeti
tive nature o f much dream content, including the nightm arish
dreams accompanying posttraum atic stress disorder and tempo
ral-lobe epileptic seizures.
■ The fact th a t dreams can be made more v iv id and frightening
by drugs affecting the dopaminergic and cholinergic systems
suggests th a t the relationship between the neural netw ork for
dreaming and dream content can be studied by determ ining the
influence o f various drugs on specific aspects o f dream content.
■ Because defects in the neural netw ork for dreaming can lead
to changes in dream content, the general relationship between
this netw ork and specific aspects o f dream content can be
5
studied by exam ining the dream reports o f patients in a wide
variety o f disease states.
■ The many established parallels between waking cognition and
dreaming raise the possibility th a t some dreams may make use
o f the system o f figurative th in kin g , w hich cognitive scientists
have shown to be pervasive in waking thought (Fauconnier,
1997; Gibbs, 1994; Lakoff, 1987).
A lth o u g h the model explains the neural and cognitive bases fo r dream-
ing and shows th a t dream content is psychologically m eaningful in th a t it is
coherent, relates to other psychological variables, and is generally continuous
w ith waking conceptions and concerns, it does n o t claim any purpose or
fu n ctio n fo r dreams. O n the basis o f current evidence, it is more lik e ly th a t
dreams are an accidental by-product o f two great evolutionary adaptations,
sleep and consciousness (Antrobus, 1993; Flanagan, 1995; Foulkes, 1993).
This p o in t is discussed in detail in chapter 6 as part o f the critique o f
tra d itio n a l dream theories.
However, the fact th a t dreams sometimes dramatize em otional preoccu
pations, or contain parallels w ith the figurative dimensions o f waking
thought, may explain w hy many societies have invented uses fo r dreams,
usually in conjunction w ith religious ceremonies or m edicinal practices.
These possible uses are n o t discussed in this book because the focus here
is exclusively on understanding the process o f dreaming and the nature o f
dream content. It therefore does not present the many anthropological
studies o f beliefs about dreams in different cultures (Tedlock, 1991). The
ric h literature on the W estern use o f dreams in re lig io n (Bulkeley, 2001)
and a rtistic creativity (Barrett, 2001) is n o t discussed. N o r is the literature
on dreams and psychotherapy considered, although systematic studies o f a
cognitive—experiential approach to the use o f dreams in therapy shows th a t
ta lkin g about them is rated as helpful by clients (H ill, 1996; H ill et al.,
2001). Thus, the book is s tric tly concerned w ith the development o f a
neurocognitive model o f dreams.
C hapter 1 presents the m ain findings on w hich the new model is based
along w ith several examples o f how the three different types o f findings can
be related to each other. The chapter also discusses the m ethodology on
w hich the developm ental findings are based in greater detail than otherwise
would be necessary, because parts o f th a t m ethodology have been inade
quately characterized in some sources. Chapter 2 explores the methods th a t
can be used to lin k dream content to the neural netw ork fo r dreaming, on
the one hand, and to waking personal concerns, on the other. I t argues
th a t good methods fo r the systematic study o f dream content do exist, even
though most o f the literature on dreams is weak as a result o f the use o f

unreliable and unvalidated coding systems, poorly collected samples, small
sample sizes, and inadequate statistical procedures.
Chapter 3 turns to a detailed presentation o f the updated version o f
the coding system developed by C alvin S. H a ll and Robert Van de Castle,
w hich is now equipped w ith new statistical procedures as w ell as a spreadsheet
fo r calculating p values and effect sizes. The chapter suggests that the system
can yield the kind o f detailed findings that would be necessary to test aspects
o f the neurocognitive model. Chapter 4 introduces a new resource for
studying dream content, DreamBank.net (http://www.dream bank.net), an
In te rn e t archive o f more than 11,000 dream reports for use in a wide range
o f studies. The chapter explains the capabilities o f the search program and
statistical procedures available on the site. Several small studies are presented
as examples o f how the site can be used to develop the neurocognitive
model, including the possibility th a t it can be used to study figurative
th in k in g in dreams.
Chapter 5 argues th a t the understanding o f dream meaning can be
advanced through detailed analyses o f long dream journals, and it presents
a methodology for doing such studies in conjunction w ith the H a ll-V a n
de Castle coding system and the search program on DreamBank.net. It
presents the basic findings from a dream journal containing 3,116 dream
reports w ritte n down over a 20-year period to demonstrate the power o f
this approach, and to show the considerable regularity and psychological
coherence th a t is present in a significant p o rtio n o f dream life . To elaborate
on the neurocognitive model, the chapter suggests th a t future studies need
to be conducted w ith participants who have experienced various types o f
neurological injuries or developmental irregularities.
As noted, chapter 6 contains a detailed discussion o f the m ain dream
theories o f the 20th century. I t begins w ith a focus on the many claims
made by Freud and Jung and shows that the em pirical literature does not
support any o f their specific hypotheses. S im ilarly, it demonstrates the failings
o f the original version o f activation-synthesis theory w hile noting that the
recent version o f the theory has taken steps in a neurocognitive direction
in the lig h t o f new findings. However, activation—synthesis theory s till lacks
a fu lly developed cognitive dimension and has little or nothing to say about
dream content. The chapter also includes an analysis o f the difficulties th a t
face a ll problem -solving theories o f dreaming. In effect, then, chapter 6
clears the way for the new research directions presented throughout the book.
In w ritin g this book, I have attempted to be as eclectic and encompass
ing as possible in incorporating ideas, methods, and findings from many
different sources. I do, however, state the weaknesses o f other explanatory
systems quite frankly, because o f th e ir amazing persistence despite the ab
sence o f systematic em pirical evidence for th e ir m ain claims about the
INTRODUCTION: THE NEUROCOGNITIVE APPROACH TO DREAMS 7

construction and meaning o f dreams. It is tim e fo r the study o f dreams to
be incorporated in to cognitive science in general and cognitive psychology
in particular if understanding is to advance in this field, w hich has been
slow to develop any new th in k in g u n til the past 10 years.
Thanks to the new methods and findings th a t emerged in the 1990s,
the d iffic u lty in advancing the scientific study o f dreaming and dreams no
longer lies in the planning and execution o f solid studies. Instead, tw o new
problems confront researchers. The first is to loosen the hold o f dubious
tra d itio n a l theories on young researchers who develop an interest in dreams.
The second is to encourage established cognitive psychologists and neuropsy-
chologists to allow th e ir students to incorporate dreams in to scientific psy-
chology’s research agenda. I hope th a t this book contributes to the resolution
o f these tw o problems.

1
TOWARD A NEUROCOGNITIVE
MODEL OF DREAMS
The neurocognitive model o f dreaming and dreams proposed in this

chapter has three basic components. First, the neurophysiological substrate
underlies and activates the process o f dreaming. Second, the conceptual
system o f schemata and scripts generates the process o f dreaming. T h ird ,
dream content results from this cognitive process. This chapter discusses each
o f these components and suggests some o f the specific ways in w hich they
may relate to each other.
TH E N E U R A L N E TW O R K FOR D R E A M IN G
Research in neuropsychology and neuroimaging converged in the late

1990s to suggest the broad outlines o f a neural netw ork for dreaming that
is very different from w hat had been imagined in the past. Some disparities
exist between the lesion and imaging results, and a few disagreements can
be found among the imaging studies, a ll o f w hich use positron emission
tomography (PET) scan technology, but the findings are striking ly consistent
and provide the starting p o in t for more detailed studies. However, it would
be extremely premature to overspecify the netw ork at a tim e when new
discoveries are being made each year and so much remains to be learned
(Antrobus, 2000a; M orrison &. Sanford, 2000).
The neural substrate for dreaming provides the necessary level o f
activation needed for dreaming and, perhaps, constrains the types o f th in kin g
th a t are possible. I t seems to be responsible for determ ining the degree o f
vividness and intensity experienced in dreaming, and it may account for
other form al features o f dreaming, such as its realistic and self'participatory
nature, its general lack o f self'reflectiveness, and its occasional incongruities
o f form . However, the neural substrate cannot account for the narrative
nature o f dreaming or the substance o f dream content, w hich are products
o f the conceptual systems discussed later in this chapter.
9
Researchers w ith varied theoretical perspectives have four m ajor areas
o f agreement about the contours o f the neural netw ork. First, the mechanisms
th a t generate rapid eye movement (REM ) sleep “ support our most v iv id
and elaborate dreaming” (Foulkes, 1999, p. 6). Second, forebrain controls
o f the REM generator are located in the tegm ental region in the m iddle o f
the pons. T h ird , a com plex forebrain netw ork is necessary for dreaming.
Fourth, this forebrain netw ork plays the m ajor role in shaping dream content
(Hobson et al., 2000b; Solms, 2000). Based on these agreements, this book
concentrates on the developm ent o f a neurocognitive model o f dreams th a t
can encompass the whole range o f dream content and relate th a t content
to waking conceptions and concerns.
Theorists also disagree on some points. For example, there are varying
opinions concerning the m ixture o f neurochemicals th a t modulate the brain
during REM sleep (Gottesm ann, 2000; Hobson et al., 2000b; Perry & Piggott,
2000). Differences also exist as to whether the neural netw ork fo r dreaming
always includes the area in the pons th a t is necessary fo r REM (Foulkes,
1999; Hobson et al., 2000b; Solms, 2000). These and various other unre
solved issues related to the neural substrate for dreaming are discussed in
chapter 6 as part o f the critiqu e o f activation-synthesis theory.
The neuropsychological and neuroimaging results are interchangeable
in some ways, but the neuropsychological studies provide the best starting
p o in t because they always include a crucial psychological component, the
presence or absence o f the subjective sense o f dreaming. The neuroim aging
studies, in contrast, are a mapping o f sleep stages, although one research
group did collect dream reports from several REM awakenings as w ell as
one non-REM (N R E M ) awakening (M aquet, 2000, p. 224). Dreaming is
hig h ly correlated w ith REM, but sleep stages are an im perfect ind ica to r o f
dreaming because at least some degree o f dreaming occurs in N REM (Foulkes,
1966, 1985; Hobson, Pace-Schott, &. Stickgold, 2000a). To connect neuro
imaging w ork more closely to the process o f dreaming, there is a need fo r
studies o f NREM periods from w hich dreams are reported and o f REM
periods in w hich no dreams are recalled upon awakening (M aquet, 2000,
p. 224). It would be especially useful to have studies o f the lig h t stage o f
N R EM (Stage II N R E M ) after the first four REM periods because many
dreams seem to occur at this tim e (Antrobus, Kondo, & Reinsel, 1995;
Cicogna, N atale, O cchionero, & Bosinelli, 1998).
The prim ary source o f neuropsychological inform ation on dreaming is
a study by Solms (1997) in w hich 361 consecutive patients w ith neurological
problems were asked in great detail between 1985 and 1989 about any
changes they had noticed in the frequency and nature o f th e ir dreaming
since th e ir in ju ry or illness. Solms then integrated the results w ith the
findings from 73 published studies in the neurological literature that m ention
deficits and excesses in dreaming. Tw enty-nine o f the 361 patients turned
out to be free o f any brain lesions. They were used as a control group because
they had been faced w ith the possibility o f brain injuries, adm itted to the
hospital, and subjected to the same routines and tests as the patients who
did suffer lesions.
The responses from the rem aining patients concerning changes in th e ir
dreaming were correlated w ith the findings from th e ir neurological tests and
brain scans. Solms then focused on the patients w ith focal brain lesions so
th a t causal inferences about specific regions o f the brain could be made.
These analyses led to the conclusion that two different types o f dreaming
“ deficits” can occur— loss o f visual dreaming and complete loss (i.e., cessa
tio n ) o f dreaming. Tw o types o f dreaming “ excesses” occur— the intrusion
o f dreaming in to waking thought and increased nightm are frequency. It is
notew orthy that a ll four types o f changes in dreaming correlate w ith waking
cognitive defects. In addition, they relate to relatively specific brain sites.
The result is the general outline o f a neural netw ork fo r dreaming th a t can
be linked at many points to waking cognition, on the one side, and to the
results o f neuroimaging studies, on the other (Solms, 1997, 2000). Figure
1.1 presents an overview o f this network.
Figure 1.1. Parts of the brain that relate to dreaming, as determined by lesion studies.
Injuries to these sites can cause defects in visual imagery in dreams (site 1), loss of
dreaming (sites 2 and 3), excessive or intrusive dreaming (site 4), increased frequency
of nightmares (site 5), and loss of REM sleep and possible loss of the activation
necessary for dreaming (site 6).
TOWARD A NEUROCOGN1TNE MODEL OF DREAMS 11

The Solms study provides seven specific findings relating dreaming
and neurological structures. First, 200 o f the 332 patients w ith brain lesions
reported no changes in dreaming. This is highly useful inform ation because
it reveals the parts o f the brain th a t are n o t necessary for dreaming. Instead
o f the diffuse cortical a ctiv ity suggested by EEG recordings using scalp
electrodes, the neural netw ork for dreaming is surprisingly localized and does
n o t include vast areas o f the brain th a t are essential to waking cognition— the
dorsolateral prefrontal cortex, the sensorimotor cortex, and the prim ary
visual cortex. For example, lesions in the dorsolateral prefrontal cortex that
cause waking deficits o f self-m onitoring and decision m aking have no effect
on dreaming. These findings are supported and supplemented by a neuro-
imaging study revealing th a t a ll o f these areas, along w ith the opercular
cortex and posterior cingulate cortex, are as inactive during REM as they
are during NREM (Braun et al., 1997).
Second, Solms found changes in dreaming due to injuries in the medial
occipito-tem poral region o f the visual association cortex in two patients.
One patient lost a ll visual imagery in dreams for a short tim e, and the other
was able to see static dream images from tim e to tim e. B oth had highly
specific lesion sites and deficits in waking m ental imagery. These findings
correspond w ith 13 cases th a t go back to the 1880s in the neurological
literature, including cases o f losses o f facial imagery or color vision in
dreaming. They also parallel findings in sleep laboratory studies by Foulkes
and his colleagues. The case w ith no visual imagery in dreams is sim ilar to
Kerr, Foulkes, and Jurkovic’s (1978) laboratory study o f a patient w ith
damage in her visual association cortex. This patient had neither waking
m ental imagery nor any visual imagery in her dreams. As to Solms’s case
w ith static dream imagery, it is described by him (1997, p. 105) as “ strikingly
rem iniscent” o f a second patient studied in the laboratory by Kerr and
Foulkes (1981).
These findings are o f theoretical interest because they show th a t the
neural netw ork for dreaming has considerable specificity. They are also of
interest because the visual impairments in dreaming have parallel waking
deficits, a finding th a t suggests th a t the netw ork has close relationships
w ith at least some aspects o f waking m ental imagery. Once again, the
neuroimaging findings are consistent w ith the neuropsychological findings,
because the visual association cortex— and the auditory association cortex
as w ell— are reactivated during REM (Braun et al., 1998).
T h ird , Solms found 47 patients w ith either unilateral or bilateral in ju r
ies in or near the region o f the parieto-tem poral-occipital ju n ctio n who
reported complete loss o f dreaming and showed a decline in waking visuospa-
tia l abilities. This discovery led to the hypothesis th a t the co rtica l netw ork
fo r spatial representation, w hich is centered in the in fe rio r parietal lobes
and im portant in the creation o f waking m ental imagery, is essential for
dreaming. Solms also reports there is even some evidence th a t the le ft
parietal region “ contributes symbolic (quasispatial) mechanisms to the dream
process whereas the rig h t parietal region contributes concrete spatial mecha
nisms,” but he also stresses th a t this claim needs further investigation (1997,
p. 271). These results, w hich are supported by many individual cases in
the literature, are also consistent w ith neuroimaging findings th a t show
reactivation o f the rig h t inferior parietal lobe, an area thought to be im por
ta n t in spatial cognition (Maquet, 2000; Maquet et al., 1996). As shown
later in the chapter, these findings also may provide a crucial lin k to develop
m ental studies o f dreaming.
A n y claim by a person th a t dreaming has been lost raises the possibility
th a t it may be memory for the dream th a t has failed. Evidence th a t such
people actually have ceased to dream comes first from the fact th a t those
who reported loss o f dreaming were no more like ly to have memory disorders
than those who reported th a t they continued to dream (Solms, 1997, pp.
160-161). Second, in two laboratory studies o f people w ith neurological
deficits, awakenings from REM did n o t produce any dream recall in most
participants. In the first study, nine leucotomized people w ith schizophrenia
who claimed th a t they no longer dreamed were awakened from a ll REM
periods during two nights in the laboratory and compared w ith a control
sample o f hospitalized patients w ith schizophrenia who had n o t been leuco
tomized (Jus et al., 1973). O ut o f 66 awakenings, only two produced dream
reports, and they were both from the same person; in contrast, h a lf or more
o f the awakenings w ith the control group led to dream reports. In the second
study, only 3 o f 12 patients who reported th a t they had not dreamed over
the course o f a 10-day observation period could recall a dream from REM
awakenings. In comparison, 7 patients w ith neurological deficits who said
they continued to dream demonstrated 75% dream recall after REM awaken
ings (M urri, Massetani, Siciliano, & Arena, 1985). It therefore seems safe
to conclude that claims concerning the loss o f dreaming are credible.
Fourth, Solms found that patients w ith b ifro ntal lesions in the w hite
m atter inferior to the fro n ta l horns o f the lateral ventricles in the ventrome-
sial region also reported loss o f dreaming. This area provides a crucial lin k
between the basal forebrain and lim b ic structures on one side and many
parts o f the fro n ta l cortex. The evidence for the importance o f this area is
based on only nine cases, but the finding is strengthened by the results
from studies o f leucotomized people w ith schizophrenia th a t were previously
overlooked in the dream research literature. The studies reported the loss
o f dreaming in 70% to 90% o f the hundreds o f people w ith schizophrenia
who were leucotomized between 1940 and 1975 as a way to control th e ir
symptoms (Frank, 1946, 1950; Solms, 1997, 2000). Moreover; as just noted,
the absence o f dreaming in people who have been leucotomized has been
confirmed w ith awakenings in the laboratory during REM (Jus et al., 1973).
TOWARD A NEUROCOGN1T1VE MODEL OF DREAMS 13

Once again h ig h lig h tin g the parallels between waking cognition and dream
ing, most o f the people in those studies were lacking in in itia tiv e , curiosity,
and fantasy in waking life. These findings also fit w ith the neuroim aging
studies, w hich show th a t the basal forebrain and lim bic region are highly
active during REM (Braun e ta l., 1997;Maquet, 2000; Nofzinger et al., 1997).
F ifth , 10 o f Solms’s patients reported an increased frequency and v iv id
ness o f dreaming, often accompanied by the intrusion o f dreaming in to
waking life. They also said th a t th e ir dreams now seemed more realistic.
The most frequently involved areas were the medial prefrontal cortex, the
anterior cingulate cortex, and the basal forebrain. Some o f the patients said
th a t they fe lt like they were always dreaming or that th e ir thoughts quickly
turned in to pictures or realistic events, a finding that suggested to Solms
(1997, pp. 198-199), follow ing an idea presented by Luria (1973), th a t they
had lost the “ selectivity o f m ental processes.” Observations by members o f
the hospital staff support the idea th a t the patients were suffering from a
confusion between dreaming and waking thought. W h itty and Lew in (1957)
reported several sim ilar cases. Damasio, Graff-Radford, Eslinger, Damasio,
and Kassell (1985) described th e ir patients w ith sim ilar lesions as suffering
from “waking dreams” (p. 269). It is also notew orthy th a t the m edial prefron
ta l cortex is involved in processes o f arousal and a ttention in waking life
and th a t injuries to this area can lead to confabulation and compulsive
fabrication (Braun et al., 1997; Hobson et al., 2000b, p. 808).
Neuroim aging studies reveal that the medial prefrontal cortex, anterior
cingulate cortex, and basal forebrain are reactivated during REM (Braun et
al., 1997; Maquet et al., 1996; N ofzinger et al., 1997). A dding a new
dimension to the picture, Nofzinger et al. (1999, 2001) discovered th a t
people w ith depression have much weaker reactivation o f the m edial prefron
ta l cortex, anterior cingulate cortex, and rig h t anterior insula in REM than
norm al participants do. It is as though th e ir neural netw ork fo r dreaming
has shrunk to a small core area. Because evidence indicates that people w ith
depression may dream less than th e ir nondepressed counterparts (Arm itage,
Rochlen, Fitch, T rive di, & Rush, 1995) and th a t the few dreams they do
have are bland (B arrett & Loeffler, 1992; Kramer & Roth, 1973), this finding
demonstrates the pote n tial o f using atypical cases to learn more about the
relationship between dream content and the functioning o f the neural
substrate fo r dreaming.
S ixth , Solms found th a t injuries to the temporal lobe caused increased
nightmares o f a repetitive nature for nine patients, five o f whom had symp
toms o f epilepsy. In keeping w ith this discovery, there are many instances
in the literature o f epileptics reporting nightmares, w hich are often caused
by tem poral-lobe seizures during N REM . These patients sometimes suffer
from daytime hallucinations as w ell (see Solms, 2000, p. 847, for a summary
and references). Then, too, studies using stereotaxic electrodes to locate
the sites causing seizures show th a t the “dreamy state” sometimes experienced
as part o f the diagnostic process is related to the tem poral-lim bic region.
In one such study, the amygdala, anterior hippocampus, and tem poral cortex
were involved in every spontaneous occurrence o f this state during the
procedure (Bancaud, Brunet-Bourgin, Chauvel, & Halgren, 1994). Thus,
the possibility arises th a t the seizures may be activating the neural substrate
for dreaming (Hobson et ah, 2000a, p. 1031; Solms, 1997, p. 243). In
addition, the neuroimaging results are consistent w ith the inclusion o f the
tem poral lobe in the neural netw ork for dreaming. They show that the
occipital-tem poral region is reactivated during REM, along w ith nearby
lim b ic areas (Braun et al., 1997; Maquet, 2000).
Seventh, and finally, Solms had 53 patients w ith brain stem lesions who
were able to state whether they continued to dream. Forty-three “ reported a
preservation o f the subjective experience o f dreaming” (Solms, 1997, p.
154). Solms concluded that the results showed th a t REM is n o t necessary
for dreaming. However, his findings are n o t fu lly convincing on this issue
because it is d iffic u lt to elim inate REM, even w ith experim ental lesions in
animals (Hobson, Stickgold, & Pace-Schott, 1998). Because Solms’s patients
were not tested in the sleep laboratory for the absence o f REM, it is therefore
n o t certain th a t the relevant areas o f the brain stem were affected. Further
more, it may be that “ any lesion capable o f destroying the pontine REM
sleep generator mechanism would have to be so extensive as to elim inate
consciousness altogether” (Hobson et ah, 1998, p. R10). S till, some reports
describe people who lost REM and remained sentient (G iro ne ll, Calzada,
Sagales, & Barraquer-Bordas, 1995; Lavie, 1984, 1990); unfortunately, they
were not awakened to see whether they could report dreams, but th e ir
positive case histories show th a t future studies o f the presence or absence
o f dreaming in people w itho u t REM may be possible.
As expected on the basis o f earlier lesion studies w ith cats (Hobson,
M cCarley, & W yzin ski, 1975; M cCarley & Hobson, 1975), the neuroimaging
studies found th a t the pontine tegmentum is far more active in REM than
NREM ; one study suggested th a t it is even more active in REM than in
waking (Braun et al., 1997). This reactivation seems to extend to the
thalamus through cholinergic pathways, and then to the basal ganglia, basal
forebrain, and lim b ic and paralim bic regions.
These findings, w hile far more specific than anything th a t could have
been imagined in the early 1990s, nonetheless leave open many questions
about the exact contours o f the dream network; the studies have m ethodolog
ical differences, and the participants had individual differences (Hobson et
al., 2000b; Maquet, 2000). They also lead to inevitable speculation about
the functioning o f the network. Braun et al. (1997) stressed th a t the activa
tio n levels during REM are comparable to those in waking but w itho u t the
involvem ent o f the frontal areas so im portant to waking cognition: “ REM
TOWARD A NEUROCOGN1TNE MODEL OF DREAMS 15

sleep may constitute a state o f generalized brain a c tiv ity w ith the specific
exclusion o f executive systems that norm ally participate in the highest order
analysis and integration o f neural inform ation” (p. 1190). Braun et al. also
emphasized the functional connections among the pons, basal forebrain,
lim b ic structures, and medial prefrontal cortex during REM.
A ccording to Maquet (2000, p. 222), the amygdala may be the central
structure in the m odulation o f cortical a ctiv ity in REM, as evidenced by
the fact th a t it is tig h tly connected to the anterior cingulate cortex and
the in fe rio r parietal lobule, w hich are reactivated in REM, but has few
connections to the dorsolateral prefrontal cortex and parietal lobes, w hich
are relatively inactive throughout sleep. M eanwhile, Nofzinger et al. (1997,
2001) highlighted the importance o f the anterior cingulate cortex, w hich
plays a role in a ttentional states, performance m onitoring, and error detec-
tio n in waking thought.
Solms (2000) stated th a t the neuroim aging findings are generally con
sistent w ith his neuropsychological findings, but he doubted th a t the REM
generator is a necessary part o f the neural substrate for dreaming. Instead,
he argued th a t dreaming is generated by the dopaminergic system, w hich
has its origins in dopaminergic cells in the ventral tegmentum, just above
the pons, and then fans out to the amygdala, anterior cingulate gyrus, and
fro n ta l cortex. He agreed th a t the cholinergic pathways originating in the
pons are the most frequent instigators o f the necessary level o f forebrain
activation, but he asserted that dreaming occurs “ only if and when the
in itia l a ctivation stage engages the dopaminergic circuits o f the ventrom esial
forebrain” (Solms, 2000, p. 849). A ctivation-synthesis theorists, on the
other hand, see the neuroimaging results as strong support for th e ir emphasis
on the brain stem generator; at the same tim e, they welcome the insights
in to the forebrain netw ork provided by both the neuroimaging and neuropsy
chological findings (Hobson et al., 2000a, 2000b).
These slightly different perspectives share the idea th a t the association
cortices, paralim bic structures, and lim b ic structures may operate as a closed
loop to generate the process o f dreaming. T his loop is the starting p o in t
fo r the neurocognitive model proposed in this book. O n the one hand, this
subsystem is cut o ff from the prim ary sensory cortices th a t provide inform a
tio n about the external w orld and, on the other, from the prefrontal cortices
th a t integrate incom ing sensory inform ation w ith memory and em otion in
the process o f decision making (cf. Braun et al., 1998, p. 94). This model
implies th a t an unconstrained and freewheeling conceptual system can
operate when sufficient activation occurs. Its relative isolation may account
fo r the “ single-mindedness” o f dreams (i.e., the lack o f parallel thoughts
and reflective awareness; Rechtschaffen, 1978, 1997). A t the same tim e, as
evidence presented throughout this book shows, the neural netw ork for
dreaming contains enough cognitive processing areas, such as the medial
fro n ta l cortex, anterior cingulate cortex and, perhaps, the o rb ita l-fro n ta l
cortex, to produce coherent dramatizations th a t often portray the dreamer’s
conceptions and concerns in waking life (Foulkes, 1985, pp. 209-213; H a ll,
1953b). This emphasis on conceptions and concerns, w hich is based on
inferences from detailed studies o f dream content, provides the cognitive
dimension th a t is lacking in activation-synthesis theory.
A lthough the basic outlines o f the neural network fo r dreaming seem
clear, it is equally certain that much remains to be learned about its fu n ctio n
ing. In addition to further neuroimaging studies using fM R I and transcranial
magnetic stim ulation as w ell as PET scans, this process could be greatly
aided by clin ica l neuropsychologists who fam iliarize themselves w ith the
new model and then look for people w ith either pure lesions in relevant
areas or complaints about changes in th e ir dreaming. Such studies could be
especially helpful w ith people who previously kept a dream journal and are
w illin g and able to report dreams as th e ir lesions heal.
For now, w hat seems certain is th a t progress toward an increasingly
detailed mapping o f the neural netw ork for dreaming is inevitable: Neuropsy
chology is making rapid strides, and neuroimaging studies are becoming
increasingly sophisticated and common. Because o f the potential o f transcra
n ia l magnetic procedures, it m ight even be possible to shape dream content
by stim ulating different regions w ith in the neural netw ork for dreaming
(Mazziotta, Toga, & Frackowiak, 2000; Stewart, Ellison, W alsh, & Cowey,
2001). The stage is therefore set for a consideration o f how this netw ork
m ight be integrated w ith other areas o f dream research to build a neurocogni-
tive model. For example, the forebrain portion o f the neural substrate for
dreaming seems to be a good starting p o in t for understanding the occasional
occurrence o f an awareness o f dreaming during a dream.
TH E Q U ESTIO N OF “ L U C ID D R E A M IN G ”
The phenomenon o f becoming aware o f a dream w hile it is ongoing

enjoyed a flurry o f attention and speculation in the 1980s under the m orally
toned label o f lucid dreaming— im plying a superior or elite status fo r “ lucid
dreamers” — and efforts were made to lin k it to m editation and other altered
states o f consciousness (Gackenbach &. Bosveld, 1989; LaBerge, 1985).
A lth o ug h often remarked upon in books on dreams in the prelaboratory
era, lucid dreaming could not be studied systematically u n til it was shown
in the laboratory th a t it occurs during REM (LaBerge, Nagel, Dement, &.
Zarcone, 1981). A lth o ug h too few people have been studied in the laboratory
to establish the frequency o f lucid dreaming, laboratory and nonlaboratory
studies suggest th a t the degree o f self-awareness and sense o f conscious
control can vary greatly from person to person and even w ith in any given
TOWARD A NEUROCOGNITIVE MODEL OF DREAMS 17

lucid dream (Barrett, 1992). However, as Foulkes (1990b) noted, much
more laboratory work needs to be done concerning “ the conditions under
w hich certain kinds o f generic and autobiographical knowledge prove to
be accessible during dreaming in the service o f an ongoing comprehension
and evaluation o f dream events” (p. 121).
I f dreaming is the form th a t consciousness takes during sleep (Foulkes,
1999), and if changes in the neural netw ork for dreaming underlie different
dreaming states, then lucid dreaming may be the product o f a dream state
in w hich the higher order neural patterns th a t give human beings “ core
consciousness” and an “ autobiographical s e lf’ are more active than usual
(Damasio, 1999). This speculation is consistent w ith Rechtschaffen’s (1997)
use o f the confabulations caused by frontablobe injuries to argue th a t the
loss o f reflective awareness in dreams is the result o f the lack o f fro n ta l-
lobe a ctivity. It also fits w ith the finding th a t the high levels o f alpha a ctiv ity
during REM are related to lucid dream reports (O gilvie, 1982; Tyson, O gilvie,
&. H un t, 1984) and w ith the knowledge th a t self-awareness during REM is
associated w ith phasic (i.e., in te rm itte n t) activation w ith in the REM period
(Bradley, H o llifie ld , &. Foulkes, 1992). The content o f lucid dreams also
has a more realistic nature, w hich would be expected from this lin e o f
reasoning (Gackenbach, 1988).
Then, too, it is notew orthy that dream reports in an exploratory PET
scan study o f 12 male participants showed a greater sense o f control when
the m edial fro n ta l cortex was more active, and a greater sense o f things
being out o f control when the amygdala was most active (Shapiro et al.,
1995). Some nonlaboratory evidence suggests th a t the neural netw ork for
dreaming includes more fro n ta l cortex a ctiv ity during lucid dreaming than
during nonlucid dreaming: Lucid dreams seem to occur most frequently in
the home setting after an early m orning awakening— between 5:00 a.m.
and 6:30 a.m.— that is follow ed by imagery rehearsal and a conscious attem pt
to be aware o f dreaming upon fa llin g back to sleep (LaBerge, 1985). Thus,
the new question o f interest is the state o f the neural netw ork fo r dreaming
during the experience o f lucid dreaming and how th a t state relates to
indicators o f REM and Stage II N REM (Hobson et al., 2000a, p. 1020;
Hobson et al., 2000b, p. 837).
T H E DEVELO PM ENT OF D R E A M IN G C O G N IT IO N
The serendipitous discovery o f REM sleep in 1953, especially the

finding th a t the four or five REM periods o f the n ig h t occupy 20% to 25%
o f adult sleep tim e and lead to dream reports from 80% to 90% o f awakenings
in norm al adults, triggered an enormous advance in the understanding o f
both sleep and dreaming (Dem ent & K leitm an, 1957a, 1957b; Foulkes,
THE SCIENTIFIC STUDY OF DREAMS

1966; Kamiya, 1961). Those studies demonstrated th a t dreaming is far more
ubiquitous in both REM and N REM than any previous dream theorist ever
imagined, a finding th a t has m ajor im plications, discussed in chapter 6, for
tra d ition a l clin ic a l and functional theories o f dreaming. They also revealed
th a t dreaming has many im portant parallels w ith waking cognition.
In addition, laboratory dream studies show th a t dreaming cannot be
triggered by external stim uli and th a t it is d iffic u lt to influence dream content
w ith either presleep stim uli, such as fear-arousing or exciting movies, or
w ith concurrent stim u li administered during REM, such as a spray o f water,
sounds, or the names o f significant people in the dreamers’ lives (Foulkes,
1985, 1996a; Rechtsehaffen, 1978). For example, in a large study comparing
the influence o f neutral and affect-arousing presleep film s on the REM
dreams o f 24 adult participants, only 5% o f 179 awakenings showed any
sign o f incorporation (Foulkes & Rechtsehaffen, 1964); sim ilar results were
obtained in a study o f boys between ages 7 and 11 (Foulkes, P ivik, Steadman,
Spear, & Symonds, 1967). In the most frequently cited study o f the influence
o f external stim u li applied during REM, the sound o f a bell was only incorpo
rated in 20 o f the 204 instigations w ith 12 participants (Dem ent &. W olpert,
1958b). Somatosensory stim uli from sprays o f water, electrical pulses to the
hand, and pressure from a blood pressure cu ff seem to have the highest rate
o f incorporation, at around 40% (Sauvageau, Nielsen, & M ontplaisir, 1998,
p. 132). However, the criteria for incorporation were loose in most o f these
studies, and the researchers sometimes included what were assumed to be
m etaphoric expressions o f the stimulus (A rk in & Antrobus, 1991).
O verall, Foulkes’s judgm ent seems to be the best starting p o in t for
developing a neurocognitive model o f dreaming:
Probably the most general conclusion to be reached from a wide variety

of disparate stim u li employed and analyses undertaken is th a t dreams
are relatively autonomous, or “ isolated,” m ental phenomena, in th a t
they are n o t readily susceptible to either in duction or m odification by
immediate presleep m anipulation, at least those w ith in the realm o f
possibility in ethical human experim entation. (1996a, p. 614)
However, on the rare occasions when stim u li are incorporated, “ the

speed and ingenious fit o f the incorporation in to the meaning and imagery
context o f the ongoing dream” are often “remarkable,” suggesting the high
cognitive level o f the dreaming brain (Antrobus, 2000b, p. 474).
Laboratory studies further show that the content o f dream reports,
whether from REM or NREM awakenings, is in large measure a coherent
and reasonable sim ulation o f the real world. This conclusion joins w ith the
neuropsychological findings on people w ith brain lesions in suggesting that
a greater parallel exists between waking thought and dreaming than is
assumed by either clin ic a l or activation-synthesis theorists (Cavallero &

Foulkes, 1993; Foulkes, 1985; M eier, 1993; Snyder, 1970; Strauch & M eier,
1996). Some o f the laboratory findings on dream content are discussed in
detail in chapter 2.
In addition, findings from three laboratory studies suggest th a t waking
thought can have dream like qualities when participants are relaxing in a
darkened room. In the first two o f these studies, awake participants m onitored
by EEG gave dream like responses to 15% to 20% o f the requests fo r reports
o f w hat was going through th e ir minds (Foulkes & Fleisher, 1975; Foulkes
& Scott, 1973). In another laboratory study, judges who compared REM
reports w ith thought reports from awake participants reclining in a darkened
room rated the waking reports as more dreamlike (Reinsel, Antrobus, &
W ollm an, 1992; Reinsel, W ollm an, & Antrobus, 1986). Furthermore, a fie ld
study o f waking consciousness— w hich used pagers to contact participants—
discovered th a t 9% o f the 1,425 thought samples had “ more than a trace”
o f dream like thought and another 16% had a “ trace” o f such thought (K linger
& Cox, 1987/1988). Taken together, these studies lead to the idea th a t
dreaming may not always be a fu n ctio n o f sleep, thereby providing another
possible lin k between waking cognition and dreaming. Instead, at least some
forms o f dreaming sim ply may require a high level o f brain a ctiva tion in
com bination w ith a reduction in external stim ulation and a decrease in
self-control (Antrobus, 1991; Foulkes, 1999; Llinas & Pare, 1991).
W ith in the context o f this general evidence fo r the overlap o f waking
cognition and dreaming, two large-scale studies o f dreaming in children—
one longitudinal, one cross-sectional— provide systematic evidence th a t
offers a developm ental dim ension to a neurocognitive model o f dreams
(Foulkes, 1982, 1999; Foulkes et al., 1990). The longitudinal study began
w ith seven boys and seven girls ages 3 to 4 and took place over a 5-year
span; the study also included eight girls and eight boys ages 9 to 10 to
account fo r the years between ages 9 and 15. Remarkably, all 14 children
in the younger group participated in a ll 5 years o f the study. Tw elve o f the
16 children in the older group completed the study; the other 4 moved out
o f town.
T o check on the possibility that participation in the study im proved
dream recall and accounted for any increases in the frequency and narrative
com plexity o f dream reports, six boys ages 11 to 13 were added to the older
group in the th ird year and seven girls ages 7 to 9 were added to the younger
group in the fifth year. The new participants generally did not d iffer on any
dream measures from the original participants. In to ta l, 26 children between
ages 3 and 15 participated for 5 fu ll years, 34 participated for at least 3
years, and 43 participated for at least 1 complete year. N orm ative dream
data for each group were collected during the first, th ird , and fifth years o f
the study, when children slept in the laboratory for nine nights each. They
responded to three awakenings per n ig h t from either REM or N R EM , for

a to ta l o f 2,711 awakenings. A ll the awakenings were carried out by Foulkes
to ensure experim enter consistency. D uring the second and fourth years,
the children participated in a variety o f m ethodological studies, the most
im portant o f w hich compared dreams collected after a n ig h t o f uninterrupted
sleep in the laboratory w ith dreams collected in the m orning at home
by parents.
In addition to gathering inform ation on the frequency o f dream recall
and the content o f the dream reports, other members o f the project team
administered a wide range o f personality and cognitive tests and obtained
inform ation about school performance. Observations o f the social skills and
play patterns o f the youngest group were made at a 2-week nursery school
during the first three summers o f the study. A to ta l o f 657 nondream variables
were correlated w ith the dream data because “ it would have constituted
crim inal neglect to have collected so many dream data and n o t to have
searched far and wide for waking variables related to them ” (Foulkes, 1999,
p. 49).
The cross-sectional study focused on children ages 5 to 8 to see whether
the most interesting results o f the longitudinal study could be replicated.
I t included 20 children at each age who were w ith in 1 m onth o f th e ir
birthdays, so a to ta l o f 80 children spent three nights in the sleep laboratory.
They were each awakened 10 times, and a ll 800 awakenings, as in the
longitudinal study, were carried out by Foulkes. The children took several
cognitive tests measuring visuospatial, verbal, descriptive, and memory a b ili
ties that had correlated w ith dream recall or length o f dream reports in the
first study. The children also took three interview-based tests th a t claimed
to measure aspects o f the development o f self-awareness. In neither study
did Foulkes know the results o f the daytime tests u n til he had collected a ll
the dream data.
There are several replicated results from these tw o studies that are
im portant fo r a neurocognitive model o f dreams. None o f the findings on
rate o f recall, report length, or narrative com plexity showed any gender
differences. First, and most unexpected, the median rate o f dream recall was
only 20% to 30% from REM awakenings u n til ages 9 to 11, when the
median recall rate o f 79% from REM awakenings approached adult levels.
Recall from NREM awakenings w ent from 6% at ages 5 to 7 to 39% at
ages 11 to 13. For both REM and N REM awakenings, recall came first from
awakenings late in the night, then from awakenings in the m iddle o f the
night, and fin a lly from awakenings early in the sleep period.
Second, u n til ages 13 to 15, the children’s dream reports had different
content from w hat is reported by adults. For children under age 5, the REM
reports consisted prim arily o f static and bland images in w hich they saw an
anim al or were th in kin g about eating or sleeping. The dreams o f children
ages 5 to 8 showed a sequence o f events in w hich characters moved about

and interacted, but the dream narratives were n o t w ell developed. The
dreamer did n o t appear regularly as an active participant in her or his dreams
u n til around age 8. Compared w ith the dream reports o f adults, those o f
the young children were notable for th e ir low levels o f aggressions, m isfor
tunes, and negative emotions (D om hoff, 1996; Foulkes, 1982,1999). Gender
differences in dream content did begin to appear in late childhood (D om hoff,
1996; Foulkes, 1982) but were more prevalent by adolescence (T rupin,
1976). The findings on children ages 9 to 15 have been replicated and
extended in a m ajor lon g itu d in a l project by Strauch (1996; Strauch &
Lederbogen, 1999) th a t was based on 12 boys and 12 girls studied at 2-year
intervals in the sleep laboratory and at home.
The results on b o th recall and content are o f great theoretical im por
tance because they suggest th a t young children do not dream in the fashion
assumed by a ll previous theorists on the basis o f anecdotal and clin ic a l
accounts. Instead, they reveal dreaming to be a cognitive achievement th a t
develops gradually in the same way in w hich most other cognitive abilities
develop in children. The frequency and cognitive structure o f ch ild re n ’s
dreams is n o t adultlike u n til ages 9 to 11, and the dream reports are not
adultlike in length or content u n til ages 11 to 13. The content generally
differs from what was expected on the basis o f anecdotal accounts and
nonlaboratory studies.
Foulkes’s findings on the waking correlates o f dreaming and dream
content in children provide further surprises because verbal and lin g u istic
skills did n o t play a role in dream recall or report length u n til dreaming is
fu lly developed, and none o f the personality measures correlated w ith dream
content u n til preadolescence. The one good and consistent predictor o f the
frequency o f dream reporting in children ages 5 to 9 in both studies was
visuospatial skills, as best measured by the Block Design test o f the W echsler
Intelligence Scale for C hild re n (W IS C ) and the Embedded Figures Test.
T his finding leads to the hypothesis th a t visual im agination may develop
gradually and be a necessary cognitive prerequisite for dreaming.
There has been no rush to draw out the im plications o f these findings,
perhaps because they do n o t agree w ith common sense: Everyone has anec
dotal examples o f dreams from young children, and children seem to under
stand the concept o f dreaming. In addition, as many as h a lf o f college
students in two different samples claimed to remember a dream from c h ild
hood, although it may be significant that the average estimated age for such
dreams is 6.5 years, compared w ith the usual 3.5 years for th e ir earliest
memory (Dom hoff, 1993a). Skeptics therefore argue th a t the low rates o f
recall in young children may be the result o f waking cognitive factors rather
than a lack o f dreaming.
For example, H u n t (1989) thinks the problem may be an in a b ility to
distinguish the “ embedded” experience o f a dream from sim ilar subjective

states; others say th a t children sim ply may lack the linguistic skills to translate
the nonverbal experience o f dreaming in to the narrative report necessary
to show evidence o f dreaming (Hobson et al., 2000b; W einstein, Schwartz, &
A rk in , 1991). Foulkes found these alternative explanations u nlikely because
none o f the several linguistic, descriptive, memory, or storytelling tests
administered to the children correlated w ith rates o f recall. Such explana
tions are also contradicted by the fact th a t both REM and N R EM reports
are first given late in the sleep period; it does not seem lik e ly th a t either
discrim inatory or narrative skills would be unavailable earlier in the n ig h t
once they had developed.
The idea that young children do not dream w ell u n til th e ir visuospatial
skills are developed is supported by Foulkes’ unanticipated findings w ith
tw o o f the boys ages 11 to 13 who were added to the study during its th ird
year. Both boys had average memory and verbal skills and were adequate
students in school; but: they both scored low on visuospatial skills. N eith e r
boy reported m jiny dreams during REM awakenings, far below the average
fo r a ll other children in th e ir age group. Because neither boy lacked the
linguistic skills claim ed by critics to be the reason why young children do
n o t report dreams when awakened in the laboratory, it seems more lik e ly
th a t they were n o t dreaming (Foulkes, 1982, pp. 180-181, 225-226).
Findings on the presence or absence o f visual imagery in people who
lose th e ir sight through disease or accident before or after ages 5 to 7 also
support Foulkes’s argument. As is w ell known, those who become b lin d
after this c ritic a l period “ continue to be able w hile awake to conjure up
m ental images o f persons, objects, and events, and they continue to dream
in imagery” (Foulkes, 1999, p. 15). This p o in t includes visual dream images
o f people they m et after they became blind, w hich supports the generally
accepted idea th a t they have a system o f imagery independent o f perceptual
capabilities. O n the other hand, people who become b lin d before ages 5 to
7 do n o t have waking visual imagery or visual dreams.
The like lih o o d th a t preschool children do not dream often or w ell
may have im plications for an unexpected finding in studies o f how children
cpme .to understand im agination, pretense, and dreams. Several studies
suggest that by age 3, children understand m ental states and readily d is tin
guish between the real and the imaginary. However, preschool children do
less w ell on questions inquiring about dreams: “ Whereas 3- and 4-year-olds
are reported to have a sensitive understanding o f the origins o f im agination,
early work on dreams suggests th a t children o f this same age are quite
confused about th e ir origins” (W oolley, 1995, p. 195). Some 3-year-olds
also “appeared to conceive o f dreams as shared fantasies, claim ing th a t dream
content is shared between sleeping individuals” (W oolley, 1995, p. 189).
These differences are often explained by noting th a t pretense and
im agination are deliberate m ental activities th a t are facilitated by toys and
'H o w a r d a n e u r o c o g n it iv e m o d e l o f d re a m s 23
interactions w ith adults. W oolley (1995, p. 195) speculates th a t “ dream
origins are sim ply more d iffic u lt fo r children,” perhaps because dreams are
n o t w illfu l m ental states. I f Foulkes’s findings are used as a Starting p o in t,
however, this failure o f understanding may be due to a lack o f personal
experience w ith dreams. In fact, many o f the explanations fo r dreams offered
by preschool children— th a t they are shared fantasies, th a t they come from
God, or th a t they are produced by the people who appear in them — seem
to reflect w hat they are to ld by th e ir parents along w ith w hat they deduce
from storybooks. This alternative hypothesis suggests the need fo r new
research on the way in w hich children’s “ theory o f m ind” interacts w ith
w hat they learn about dreams from th e ir culture to produce possibly fabri
cated reports when they sense an expectation or pressure to describe a dream
(C eci, Bruck, & B a ttin , 2000).
Once children have the a b ility to dream, th e ir lin g u istic and descriptive
skills begin to correlate w ith the length and narrative com plexity o f th e ir
dream reports. S till, it is n o t u n til ages 11 to 13 th a t dream content shows any
relationship to personality dimensions. For example, the more in d ivid u a listic
and assertive children portray themselves as more active in th e ir dreams.
C hildren w ith more violence in th e ir waking fantasies have more aggressive
interactions in th e ir dreams, and those who display the most h o s tility before
going to bed in the laboratory more often dream o f themselves as angry.
These findings on the co n tin u ity o f dream content w ith waking thought
support findings in earlier studies o f children in the laboratory (Foulkes,
1967; Foulkes, Larson, Swanson, & Rardin, 1969; Foulkes et al., 1967) and
suggest th a t dreams can reflect personal concerns and em otional preoccupa
tions once there is an adequate level o f cognitive development. As shown
in D om hoff (1996) and evidence presented throughout this book, this finding
is a ll th a t remains o f the large claims by Freud and Jung (see chapter 6,
this volum e).
Foulkes’s findings raise the possibility th a t the development o f dreaming
may be based on the m aturation o f the neural netw ork for dreaming discussed
in the previous section. This hypothesis is the first and most crucial one in
an effort to create a neurocognitive model o f dreams. The idea is suggested
most strongly by the parallel between the dependence o f dreaming in c h il
dren on visuospatial skills, w hich are based p rim arily in the parietal lobes
(Robertson, 1998), and the loss o f dreaming in adults w ith injuries to
either parietal lobe. The hypothesis is also suggested by the static nature
o f preschool children’s dreams, w hich may relate to the absence o f movement
imagery in the dreams o f adults w ith lesions in specific areas o f the visual
association cortex.
M ore generally, if the low levels o f dreaming in children and the
differences in th e ir dream reports from the dream reports o f norm ative adult
samples are treated as though they are “ deficits,” then the search could be

made fo r possible causal “defects” in the neural netw ork necessary for dream-
ing. W elsh, Pennington, and Groisser (1993) follow ed this strategy in study
ing the development o f frontal-lobe executive functions in children; the
researchers used neuropsychological tests in conjunction w ith standard de
velopm ental tests. The search could be widened to include neuroimaging
studies o f the developing brain as w ell as m yelination studies (Chugani,
1999; Paus et al., 1999; R ivkin, 2000; Thatcher, 1996). Indeed, because
m yelination o f the inferior parietal lobules is n o t fu nctionally complete
u n til ages 5 to 7, dreaming may n o t be fu lly developed u n til after that age
period (Janowsky & Carper, 1996; M ark Solms, Departm ent o f Neurology
and Psychology, U niversity o f Cape Town, personal com m unication, June
14, 1999). The fruitfulness o f this approach is also seen in studies showing
th a t the presence or absence o f visual imagery in b lin d adults depends on
whether they lost th e ir sight before or after ages 5 to 7 (H urovitz, Dunn,
Dom hoff, & Fiss, 1999).
The integration o f the neural netw ork fo r dreaming w ith Foulkes’s
developmental findings would provide a solid basis fo r a model that is
genuinely neurocognitive instead o f simply neuropsychological, in the sense
th a t doing so could relate a neural system to the development o f dreaming.
However, it is necessary to include what is know n about the nature o f dream
content before the cognitive dim ension o f dreams can be folded in to modern-
day cognitive theory and then incorporated in to a neurocognitive model
o f dreams.
T H E N A TU R E OF D R EAM C O N T E N T
A lth o ug h there are several systems o f content analysis th a t have made

one or more contributions to the understanding o f dream content (Foulkes
& Shepherd, 1971; G ottschalk &. Gleser, 1969; W inget & Kramer, 1979),
the largest and most systematic body o f findings on what people dream
about comes from a comprehensive set o f descriptive em pirical categories
developed b y H a ll ( 1951) and then fin alized-w ith the help o f Van~de Castle-
(H a ll & Van de Castle, 1966). Four general findings w ith this H a ll-V a n
de Castle system must be encompassed by a neurocognitive model.
First, several different studies revealed th a t the dream lives o f college
men and women in the U nited States remained the same throughout the
second h a lf o f the 20th century despite m ajor cultural changes (Dom hoff,
1996; Dudley & Swank, 1990; H a ll, Dom hoff, B lick, & Weesner, 1982;
H a ll & Van de Castle, 1966; Tonay, 1990/1991). These findings also provide
a norm ative basis fo r many other studies, as explained in chapter 3. Second,
little or no change in dream content occurs once adulthood is reached.
T h a t is, older dreamers do n o t differ from college students, except perhaps

for a decline in physical aggressions and negative emotions (C ote, Lom e-
Lussier, Roy, & D eKoninck, 1996; H a ll &. Dom hoff, 1963b, 1964; Howe
& B lick, 1983; Inge Strauch, Department o f Psychology, U niversity o f
Zurich, personal com m unication, January 24, 2000; Zepelin, 1980), nor
does dream content change much in longitudinal studies o f dream journals
provided by adults, a claim that holds true for periods as long as four or
five decades and for people keeping journals in to th e ir 70s (D om hoff, 1996;
H a ll &. Nordby, 1972; Lortie-Lussier, Cote, &. Vachon, 2000; S m ith &
H a ll, 1964).
The th ird relevant result w ith the H a ll-V a n de Castle system is th a t
there is a stable pattern o f cross-cultural sim ilarities and differences in dream
content. M ost o f the research on w hich this conclusion is based, much o f
it unpublished w ork by H a ll, is summarized in D om hoff (1996, chapter 6).
N o significant additions to this literature have been made since the volum e
was published. Everywhere in the world, for example, women and men have
the same differences in the percentage o f gendered characters who are men
or women. W om en dream equally o f men and women, but 67% o f the
gendered characters in m en’s dreams are other men (H a ll, 1984). The same
gender differences are found in short stories by male and female authors
(H a ll &. Dom hoff, 1963a) and in stories to ld by preschool children (D om hoff,
1996, p. 89). The low percentage o f men in the dreams o f Japanese women
seemed to be the m ajor exception to this generalization (Yamanaka, M orita,
& Matsumoto, 1982), but a later study o f several different samples revealed
the same percentages as elsewhere (Nishigawa, Brubaker, & Dom hoff, 2001).
For both men and women cross-culturally, dreams usually contain more
aggression than friendliness, more misfortune than good fortune, and more
negative emotions than positive emotions. In addition to these sim ilarities,
there are also a few differences that make sense on the basis o f cultural
differences. For instance, people in small, tra d itio n a l societies have a higher
percentage o f anim al characters in th e ir dreams than people from large,
industrial societies. M oreover, there are large variations from society to
society in the percentage o f a ll aggressive interactions th a t are physical in
nature, although men in most societies have a higher percentage o f physical
aggression in th e ir dreams than women do (D om hoff, 1996; Gregor, 1981;
O ’N e il & O ’N e ll, 1977).
Finally, studies o f dream journals have demonstrated wide ind ividu a l
differences on a variety o f H a ll-V a n de Castle content indicators, w hich
are explained in chapter 3. Those differences generally relate to the waking
concerns or past em otional preoccupations o f the dreamers. Thus, a co n tinu
ity exists between most aspects o f dream content and waking thought (B ell
&. H a ll, 1971; Dom hoff, 1996; H a ll & Lind, 1970; H a ll &. Nordby, 1972).
This finding leads to the hypothesis o f a continuity principle, w hich is com pati-

ble w ith Foulkes’s (1967, 1982, 1999) findings in laboratory studies w ith
both children and adults.
The co n tin u ity principle is best demonstrated by b lin d analyses o f
dream journals, in w hich nothing is know n about the dreamer u n til he or
she later answers questions developed on the basis o f the content analysis.
In particular, b lin d analyses lead to accurate portrayals o f the dreamers’
conceptions and concerns regarding the im portant people in th e ir lives.
This emphasis on questions developed from the results o f content analyses
follows from three conclusions th a t are based on earlier attempts to find
correlations between dream content and standard personality measures.
First, the findings w ith projective techniques are meager and inconsistent
(Dom hoff, 1996; H a ll, 1956), a situation th a t may be a fu n ctio n o f the
inadequacies o f those instruments (L ilien feld , W ood, &. Garb, 2000). Sec
ond, the results w ith structured personality tests, although usually consistent
w ith the co n tin u ity principle, did n o t lead to new insights, so such tests
were seldom used after the early 1970s (D om hoff, 1996). T h ird , past research
shows th a t dreams most directly reveal concerns, interests, and worries,
rather than personality traits, suggesting th a t an open-ended neurocognitive
approach may be most useful at this juncture (Dom hoff, 1996, chapter 8;
H all, 1953c; H a ll & Nordby, 1972). Chapter 5 explains how studies using
questions th a t are based on a dream series are conducted.
Several of the discoveries w ith the H a ll-V a n de Castle system, espe
cially the consistency o f adult dream content over tim e, lead to the idea
th a t a repetition principle operates in the dream process at least some o f the
tim e (Dom hoff, 1993b, 1996). The tendency to repeat has gone unnoticed
by those who study one dream at a tim e w ith clin ica l study participants,
use samples o f individual dream reports from groups o f people, or hold to
Jung’s (1974) theory th a t a dream series shows a pattern of symbolic change
toward greater personal integration. The relative absence o f the repeated
themes in dreams collected over several weeks from participants in laboratory
studies suggests th a t the pervasiveness o f repetitive dreaming may be overesti
mated by selective dream recall in everyday dream journals (D avid Foulkes,
personal com m unication, M arch 25, 2001). Thus, the fact o f repetition is
solidly established, but its relative frequency remains to be determined.
The idea o f a repetition principle in dreams not only describes the
consistency over years and decades in characters, social interactions, a ctivi
ties, and settings in the longitudinal studies using the H a ll-V a n de Castle
system but also encompasses three other repetitive aspects o f dream life that
must be comprehended w ith in a neurocognitive model o f dreaming. First,
the extensive clin ic a l literature on the repetitive nightmares o f people
suffering from posttraum atic stress disorder fits w ell w ith the idea o f a
repetition principle (Hartm ann, 1984, 1998; Kramer, 2000b; Kramer,

Schoen, & Kinney, 1987). This literature shows th a t such dreams are more
frequent and persistent than was realized u n til systematic studies began in
the afterm ath o f the V ietnam W ar (Barrett, 1996).
Second, the re p etitio n principle can encompass the recurrent dreams
th a t 50% to 80% o f people claim to have had at one tim e or another in
th e ir lives. Such dreams often start in late childhood or early adolescence,
sometimes last fo r a lifetim e, and are usually hig h ly negative in content
and em otionally upsetting (C artw right & Romanek, 1978; Dom hoff, 1996;
Zadra, 1996). T h ird , the idea o f a repetition principle can incorporate the
repeated themes found in most series o f 20 or more dreams (H a ll, 1947,
1953c). In other words, it is not just H a ll-V a n de Castle indicators th a t
are consistent over many years, but also general themes, such as being lost,
preparing meals, or being late for an exam ination. In a study o f 649 dreams
over a 50-year period, fo r example, just six themes accounted for at least
part o f the content in 71% o f the dream reports (Dom hoff, 1993b).
The concept o f a repetition principle suggests several potential links
between dream content and the neural substrate fo r dreaming, particularly
dreams’ possible relationship to the vigilance-fear system th a t seems to be
centered in the amygdala (LeDoux, 1996; W halen, 1998). The best examples
o f this point, o f course, are the repetitive nightmares o f posttraum atic stress
disorder. These nightmares sometimes happen in Stage II o f N R EM (Van
der K olk, Blitz, Burr, Sherry, & H artm ann, 1984) and seem to have parallels
w ith the nightmares people w ith epilepsy suffer as a result o f seizures in
N R EM (Solms, 1997, 2000). In addition, as noted in the discussion o f the
neural network for dreaming, the dreamy states sometimes experienced by
epileptics are usually related to the tem poral-lim bic region (Bancaud et al.,
1994). Thus, future neuroim aging work on both posttraum atic stress disorder
and epilepsy may hold promise for links between the repetition principle
and the neural netw ork for dreaming.
However, there need n o t be an exclusive focus on people w ith brain
injuries. The consistency o f em otionally painful themes and o f heightened
scores on H a ll-V a n de Castle indicators in the dreams o f many people
w ith o u t diagnoses o f brain in ju ry or other m edical conditions suggests th a t
th e ir dream life is often “ stuck” in the past in a way that fits w ith the
persistence o f negative memories stored in the vigilance-fear system
(D om hoff, 1996; see chapter 5). B oth dreams and the vigilance—fear system
seem to provide a neurocognitive record o f traumas, upsets, and tensions over
a lifetim e. Moreover, b o th may persist even when the person is em otionally
recovered and unhampered by the past during waking life. T his possibility
suggests that dreams may n o t always be sym ptom atic o f present-day problems,
contrary to what a ll c lin ic a l theories assume.
Systematic studies showing the effects o f different drugs on dream
content, when conducted in conjunction w ith neuroimaging studies, m ight

help p in p o in t relationships between re p e titive dream co n te n t and specific
components o f th e dream -generation netw ork. T he prom ise o f such studies is
seen in the fa ct th a t b o th the a n tich o lin e rg ic beladonna alkaloids (K etchum ,
S idell, C row ell, A ghajanian, & Hayes, 1973; W ic h lin s k i, 2000) and dopa
m ine (H artm ann, Russ, O ld fie ld , Falke, & Skoff, 1980; Solms, 2000) in te n
sify the dream experience. People w ith epilepsy or Parkinson’s disease m ig h t
be candidates fo r such content studies because it already is kn o w n th a t
the m edications th a t elim ina te e p ile p tic seizures also reduce or elim in a te
nightm ares and th a t L-dopa potentiates the dream experience fo r people
w ith Parkinson’s disease (H artm ann, 1984; Perry, W alke r, Grace, & Perry,
1999; Solms, 1997).
A lth o u g h early studies concerning the effect o f drugs on dream co n te n t
led to few clear results fo r a va rie ty o f reasons (R oth, Kram er, & Salis,
1979), the p o te n tia l fo r pretest—posttest studies o f in d iv id u a l cases is shown
in the large positive changes in the dream con te n t o f a 21-year-old wom an
afte r she began ta kin g sertraline (Z o lo ft), a selective serotonin reuptake
in h ib ito r, to cope w ith anxiety attacks (K irschner, 1999). The positive
changes included increased frie n d ly in teractio ns and fewer aggressive in te r
actions and negative em otions. I t is also o f interest th a t she showed a decline
in “ elements fro m the past,” w h ich m ig h t be an in d ic a tio n th a t the re p e titio n
p rin cip le was ha vin g less influence on her dreams.
Dream co n te n t and the neural netw ork fo r dream ing also m ig h t be
lin ke d by investigations th a t correlate specific neurological defects w ith
atypical scores on the H a ll-V a n de Castle indicators. People w ho have
suffered damage to the amygdala m ig h t be ideal candidates fo r futu re defect
studies because they have lost th e ir capacity fo r fear in w aking life and
express p redo m inantly positive em otions (A dolphs & Damasio, 1998; D a
mask), 1999; Pace-Schott, 2000). I t therefore could be hypothesized on the
basis o f the c o n tin u ity p rin cip le th a t the p ro p o rtio n o f negative em otions
in th e ir dreams w ould be far low er th a n the 80% figure th a t several d iffe re n t
studies have found (H a ll et al., 1982; H a ll & V an de Castle, 1966; Roussy,
Raymond, & D e K o n in ck, 2000; Tonay, 1990/1991).
The p o te n tia l fo r such studies is dem onstrated in older reports cited
by Solms (1997) th a t show a decline in “ narrative com ple xity” in the
dream reports o f people w ith specific neurological defects through in jurie s
or operations. I t is also seen in a study showing th a t 17 m en w ith ch ro n ic
b ra in syndrome had more fa m ily members, less aggression, and less em o tio n a l
con te n t in the 31 dreams they reported than does the H a ll-V a n de Castle
norm ative sample th a t is described in chapter 3 (Kram er, R oth, & T rin d e r,
1975). T his p a tte rn o f findings suggests th a t th e ir dreams were bland, a
characterization th a t fits w ith the w aking personalities o f people w ith th a t
co n d itio n (T orda, 1969). I t m ig h t even be th a t there is a d iffe re n t p ro file
on H a ll-V a n de Castle indicators fo r each type o f defect, a p o ssib ility th a t
TOWARD A NEUROCOGNITIVE M ODEL OF DREAMS 29

was dem onstrated in a sample o f 104 dream reports from 20 m en w ith
schizophrenia (D om hoff, 1999b, p. 127)
It also could be useful to lo o k fo r changes in dream co n te n t as the
process o f dream ing returns in people who have in jurie s to one o r the o ther
parietal lobe. I t m ig h t be th a t co n te n t is sim ple and banal at first, re fle ctin g
o n ly a p a rtia l recovery. T he n, too, th is approach could be used to test the
idea that the le ft parie ta l lobe is m ore in vo lve d in sym bolic (i.e., quasispatial)
constructions and the rig h t parietal lobe in concrete, spatial constructions
(Solms, 1997, p. 271).
D R E A M C O N T E N T A N D W A K IN G C O G N IT IO N
Findings fro m the study o f dream co n te n t suggest lin ks n o t o n ly w ith

th e neural n e tw o rk fo r dream ing b u t also w ith w aking co g n itio n . In p a rtic u
la r, the c o n tin u ity p rin c ip le provides the same k in d o f strong conn ection
between dream ing and w aking co g n itio n th a t has been dem onstrated by
th e neuropsychological and developm ental evidence presented earlier in
th is chapter. T h is c o n tin u ity leads to the hypothesis th a t b o th dream ing
and waking co g n itio n deal w ith the same psychological issues to a large,
extent. T h is hypothesis provides the basis fo r lin k in g a neurocognitive m odel
o f dreams w ith w hat is know n about w aking co g n itio n .
However, as the evidence concerning the re p e titio n p rin c ip le in the
previous section indicates, the c o n tin u ity p rin c ip le does n o t operate e n tire ly
according to cu rre n t personal interests and concerns. Dream co n te n t is also
continuous in varying degrees fo r d iffe re n t people w ith past w aking concerns.
Discrepancies between cu rre n t w aking concerns and cu rre n t dream co n te n t,
such as dream ing about p a in fu l events th a t are no longer th o u g h t about in
waking life , could be used to see how the c o n tin u ity and re p e titio n prin cip le s
interact w ith each o ther to shape dream co n te n t.
The sta rtin g p o in t fo r adding a co g n itive dim ension to the m odel is
th e notion o f a conceptual system, o r system o f schemata and scripts, w h ich
is the organizational basis fo r a ll hum an know ledge and beliefs. M ost o f th is
system is th o u g h t to be unconscious, in the sense th a t it is outside conscious
awareness, b u t people can become conscious o f the system as w e ll. T he
conceptual system consists o f both e xp e rie n tia lly based and fig u ra tive con
cepts, b o th o f w h ich are processed and understood equally fast and w e ll
according to experim ental studies (G ibbs, 1994, 1999). T he conceptual
system builds on three types o f exp e rie n tia l categories— basic level, spatial
relations, and sensorimotor— w h ich in tu rn are based on b o d ily sensations
a n d interactions w ith the w o rld (L a ko ff & Johnson, 1999).
Basic-level categories arise through the in te ra c tio n o f in h e rite d neural
structures w ith patterns o f s tim u li from the enviro nm ent. T hey reflect dis-

tin ctio n s among types o f animals, such as cows, horses, and goats; types o f
social interactions, such as frie n d ly and aggressive interactions; and types
o f actions, such as w a lking and ru n n in g . Basic-level categories are most
d ire ctly distinguished from other categories in th a t a single m ental image
can represent an e n tire category, such as a “dog,” “ cat,” “ boat,” or “ car”
(M urphy & Lassaline, 1997). In a d d itio n to the large num ber o f basic-level
categories, there are also spatial-relations categories th a t are e xp e rie n tia l in
nature, such as “up,” “ dow n,” “ in fro n t of,” and “ in back of.” C om parative
lin g u is tic studies show th a t “ there is a re la tive ly sm all co lle c tio n o f p rim itiv e
image schemas [sic] th a t structure systems o f spatial relations in the w o rld ’s
languages” (L a ko ff & Johnson, 1999, p. 35). F inally, sensorim otor categories
are based on d ire ct experience o f such varied qualities as tem perature,
m otion, and touch.
Dreams are th o u g h t o f as h ig h ly “ sym bolic” in m any d iffe re n t cultures,
in clu d in g W estern c iv iliz a tio n , b u t th e findings from con te n t analysis suggest
th a t dreams may consist p rim a rily o f constructions arising fro m e xpe riential
categories. Based on his reading o f thousands o f dreams collected from
children, teenagers, and adults in the sleep laboratory, Foulkes (1985) con
cluded th a t m ost dreams are sim ulations o f real-w orld experiences. Young
a d u lt dreamers are o fte n shopping, playing sports, v is itin g w ith th e ir friends,
arguing w ith th e ir parents, w orrying about the faithfulness o f th e ir lovers, or
feeling tem pted to be u n fa ith fu l them selves. T he con te n t o f young c h ild re n ’s
dreams is usually even m ore realistic.
A lth o u g h the F la il-V a n de C astle coding system is accurately described
as em pirical and descriptive, it is n o te w o rth y th a t m ost o f its coding catego
ries are basic-level categories. T h is p o in t holds true fo r a ll the social in terac
tio n , a c tiv ity , and em otions categories and fo r m ost o f the character catego
ries. T his coding system therefore makes good th eoretical sense to the degree
th a t dreams are constructed from exp e rie n tia l categories. Perhaps th is focus
on basic-level categories also explains w hy the system can be learned and
used w ith h ig h intercoder re lia b ility by new researchers in m any d iffe re n t
countries.
The theory o f cognitive fu n c tio n in g sketched out in the previous
paragraphs provides a basis fo r adding a co g n itive dim ension to the neurocog-
n itiv e m odel because the theory fits w e ll w ith earlier w o rk on dream content
by H a ll (1953b), Foulkes (1985), A n trobus (1978, 1991), Fiss (1986), and
other dream researchers who approach the to p ic from a co g n itive perspective.
T h e m odel begins w ith the p ro p o sitio n th a t dream ing is w h at the m ature
bra in does w hen (a) the neural ne tw o rk fo r dream ing o u tlin e d earlier in
the chapter is at an adequate le v e l o f a ctiva tio n , (b) external s tim u li are
occluded, and (c) the self has been relinquished (Foulkes, 1999). T h is view
accounts fo r dream ing a t sleep onset, in REM sleep, and at tim es o f sufficient
a ctiva tio n during N R E M (A ntro bus, 2000b; A ntrobus et al., 1995; Vogel,
1991). I t also explains w hy dream ing sometimes occurs in awake particip ants
w ho are resting q u ie tly in a darkened sleep laboratory, where EEG recordings
ve rify th a t the pa rticip a n ts are in fa ct awake (Foulkes & Fleisher, 1975;
Foulkes & S cott, 1973).
O nce instigated, dream ing draws on m em ory schemata, general kn o w l
edge, and episodic m em ories to produce reasonable sim ulations o f the real
w o rld (A ntro bus, 1991; Foulkes, 1985, 1999), w ith due allow ance fo r an
occasional h ig h ly unusual o r extrem ely m em orable dream (B ulkeley, 1999;
H u n t, 1989; Knudson & M in ie r, 1999; K u ike n & S ikora, 1993). G enerally
speaking, these sim ulations express th e dream er’s “ conceptions,” w h ic h also
are the basis fo r a ctio n in th e w aking w o rld from the standpoint o f co g n itive
theory. In p a rticu la r, dreams express several key aspects o f people’s concep
tu a l systems, especially conceptions o f the self, fa m ily, and friends (H a ll,
1953b).
T he emphasis in the theory is on conceptions o f “ s e lf’ and “ others”
because studies o f ad u lt dream co n te n t show th a t dreams reflect re la tive ly
little about a person’s attitudes tow ard cu rre n t events and p o litic s (H a ll,
1951). S im ila rly, Foulkes (1982, 1999) found th a t c h ild re n between ages 5
and 15 dreamed little o f th e ir tw o m ost tim e-consum ing daytim e a ctivitie s:
going to school and w a tching television. Instead, they dreamed about recre
a tio n a l a ctivitie s. A n emphasis on the h ig h ly personal nature o f dreams
m ay exp la in w hy the dreams o f college students in the U n ite d States have
n o t changed over the past 50 years; the culture has changed, but personal
concerns probably rem ain stable. T h is emphasis also may explain w hy dreams
are m ore sim ila r th a n th e y are d iffe re n t around the w orld. A s anthropologist
Thom as G regor (1981) suggested at the conclusion o f his detailed study o f
385 dream reports from m en and wom en in a sm all n a tive group deep in
the A m azon jungle, “ it may be possible to show th a t the dream experience
is less va ria n t than o th e r aspects o f cu ltu re ” (p. 389).
S ta rtin g w ith the idea th a t dreams usually express h ig h ly personal
conceptions, it is possible to b u ild a com plex pictu re o f a dreamer’s concep
tu a l system because people usually have m ore th a n one conce ption o f them
selves and the im p o rta n t people in th e ir lives. M oreover, these conceptions
o f self and others can be co n tra d icto ry as w e ll as num erous. Some o f the
apparent co ntradictions may disappear, how ever. For exam ple, closer analysis
may show th a t a parent is seen as supportive in some contexts, such as
w hen facing exams o r problem s at w ork, b u t re strictive in others, such as
w hen the dream er wants to engage in sexual a ctivitie s (H a ll, 1947). I t is
possible th a t conceptual maps o f the dream ing m ind th a t are based on
findings from co n te n t analyses could be expressed in n e tw o rk term s (e.g.,
M arkm an, 1999; Osgood, 1959).
In ad d itio n , it is possible th a t the use o f conceptions is m ore diffuse
durin g dream ing because the cogn itive system is unconstrained by the re-

quirem ents o f the w aking w orld (Foulkes, 1985). T his hypothesis m ig h t
help account fo r the re p e titive nature o f dream co n te n t related to sig nifica nt
people and interests. I t is as though the “ updated” versions o f key concepts
are no m ore lik e ly to be used than th e older ones.
T his neurocognitive model also contains a way to assess the w e igh t
to be given to the conceptions expressed in dreams: by determ ining the
re la tive frequency o f th e ir occurrence. Because findings w ith the H a ll—V a n
de Castle system show th a t frequency reveals the in te n sity o f a concern or
interest, it can be said th a t dreams reveal b o th conceptions and concerns,
and therefore have at least some degree o f psychological m eaning. T h is
p o in t fu rth e r integrates the H a ll—V a n de Castle coding system w ith a neuro-
co g n itive m odel because n o t o n ly do its categories relate to basic-level
concepts, but its frequencies relate to conscious concerns.
Even though dreams seem to be based to a large extent on expe riential-
le ve l categories, the emphasis in a neu ro cognitive m odel on the close paral
lels between w aking th o u g h t and dream ing raises the p ossibility th a t some
o f the unusual and n o t im m ediately understandable features o f dreams may
be the product o f figurative th in k in g — conceptual m etaphors, m etonym ies,
ironies, and conceptual blends (Fauconnier, 1997; G ibbs, 1994; L a ko ff &
Johnson, 1999). Figurative concepts are sometimes th o u g h t o f as mere
em bellishm ents o f speech th a t are n o t necessary fo r th in k in g , but fo llo w in g
a wide range o f experim ental studies summarized by G ibbs (1994), m any
o f w h ich were carried o u t by him and his students, m any cogn itive scientists
now see figurative concepts as an im p o rta n t p a rt o f people’s conceptual
system. T he system o f conceptual m etaphors is learned anew by each person
as a result o f repeated experiences w ith in the course o f child h o o d
developm ent.
L a ko ff and Johnson (1999) estim ated th a t perhaps hundreds o f “p ri
m ary” conceptual m etaphors “map” w ell-understood expe riential categories
(i.e ., the source domain) to more com plex or abstract m atters o f hum an
concern (i.e., the target domain). For example, basic experiences lik e w arm th
and m o tio n are used to understand m ore d iffic u lt concepts lik e “ frie n d sh ip ”
(“ they have a warm relationship” ) and “ tim e ” (tim e often “ goes by slow ly,”
b u t sometimes “ tim e flies by” ). Just as in w aking though t, figurative th in k in g
may be used in dreams when i t expresses a conception better and m ore
succinctly than an experiential concept does (H a ll, 1953a; Lakoff, 1997).
T h is idea also provides a plausible explana tion fo r w hy m any d iffe re n t
m etaphoric expressions in dreams seem to exist fo r one “referent” : Each
m etaphor provides a s lig h tly d iffe re n t conception o f the referent object.
O ne avenue in to the p o ssibility o f a lin k between w aking figurative
though t and dream co n te n t m ight be found in ty p ic a l dreams, such as flyin g
under one’s ow n power or finding oneself in appropriately dressed in pub lic.
A co n te n t analysis o f 983 dream reports in 2-week journals kept by 126

students in a college course dem onstrates th a t fly in g dreams accounted fo r
o n ly 0.5% o f the to ta l; the figures fo r o ther typ ica l dreams— such as teeth
fa llin g out, fa llin g in space, or fin d in g m oney— are even low er (D om hoff,
1996, p. 198). H ow ever, several survey studies have suggested th a t at least
a sig n ifica n t m in o rity o f respondents have had one o r more o f such dreams
(G riffith , M iyago, & Tago, 1958; N ielsen, Zadra, G erm ain, & M o n tp la isir,
1999; W ard, Beck, &. Rascoe, 1961). These in frequ ent dreams may be
examples o f “ p rim ary” m etaphors, w h ic h are based on repeated correlations
between tw o dim ensions o f experience th a t are com m on in ch ild h o o d devel
opm ent. For exam ple, tasting som ething sweet (a physiological process) and
th e n experiencing pleasure (an em otion ) lead to the m etaphor th a t “ pleasure
is tasty” (G rady, 1999).
C onsider dreams o f fly in g under one’s ow n power, w h ich were reported
by a little m ore th a n h a lf o f college students in tw o surveys and said by
them to be generally p o sitive in tone (D om hoff, 1996). Searching fo r a
m etaphor related to flyin g , the p o ssib ility arises th a t these dreams may be
instances o f the prim ary m etaphor “happiness is up,” as found in such
expressions as “ h ig h as a k ite ,” “ w a lking on a ir,” and “ flo a tin g on cloud n in e .”
T h is speculation could e xp la in w hy people sometimes become apprehensive
about fa llin g durin g th e ir positive fly in g dreams, ju st as people w o rry th a t
they may “ crash” or “have the a ir le t o u t o f th e ir b a llo o n ” w hen th e y are
too elated in w aking life .
S im ila rly, it may be th a t dreams o f appearing in appropria tely dressed
in public, w h ic h are reportedly experienced by 40% to 50% o f college
students, usually beginning in th e ir m idteens, and sometimes m ore th a n
once, are instances o f the conceptual m etaphor “ embarrassment is exposure”
(D om hoff, 1996, p. 203). T h is m etaphor is expressed thro u g h such w e ll-
know n phrases as “ caught red-handed,” “ caught w ith egg on your face,” and
“ caught w ith your pants dow n” (H o lla n d & K ipnis, 1994). I t m ig h t be
evidence fo r th is conjecture th a t w hen college students are asked to w rite
dow n the dream in w h ich th e y experienced the greatest feeling o f embarrass
m ent, they m ost o fte n spontaneously report one in w h ich they are inade
quately a ttire d in a p u b lic place (D om hoff, 1996).
These tw o h yp o th e tica l examples aside, the few attem pts to undertake
systematic studies o f m etaphor in dreams suggest th a t m ost dreams do n o t
seem to relate obviously to prim ary m etaphors (H a ll, 1953a). R ather, m ost
dreams are lik e dramas o r plays in w h ich the dreamer acts o u t various
scenarios th a t revolve around a few basic personal themes (G reenberg &
Pearlman, 1993; H a ll, 1947). Dreams seem to be instances o f the “ th e m a tic”
p o in t on th e re p e titio n dim ension, th a t is, specific episodes or examples
re la tin g to general em otion al preoccupations, usually negative in nature.
T hey appear to take the fo rm o f proverbs o r parables, w h ich can be under-

stood only by extracting “ generic” in fo rm a tio n from specific stories (Lakoff,
1993b; L a ko ff &. Turner, 1989).
These com plex dreams may re ly on “resemblance” m etaphors, w h ich
depend on the perception o f the com m on aspects in tw o representational
schemata (G rady, 1999), o r o n conceptual blends, w h ich often start w ith
basic conceptual metaphors and th e n are elaborated in to h ig h ly nove l
thoughts (G rady, Oakley, & Coulson, 1999). H a ll (1953a) showed th a t
b lin d analyses o f a series o f dreams can lead to plausible and p o te n tia lly
verifiable inferences when figurative form s o f though t related to a m ajor
concern are used several tim es in the dream series. T o take his best example,
a young wom an w ho provided a series o f dreams had an especially s trikin g
one in w h ich she is searching fo r her wedding gown because she and her
husband are to be m arried again on th e ir first w edding anniversary. However,
she is disappointed when she found the gown: I t is d irty and to m . W ith
tears in her eyes, she puts the gown under her arm and goes to the church,
o n ly to have her husband ask w hy she has brought the gown. She reported
th a t in the dream she is “ confused and bew ildered and fe lt strange and
alone” (H a ll, 1953a, p. 179).
Looking at th e dream fro m a figurative p o in t o f view , H a ll hypothesized
th a t the state o f the dress m ig h t express her conception o f her m arriage.
In today’s terms, the dream may be a conceptual blend th a t is based on a
m etonym y. T o test this hypothesis, H a ll looked to see i f other dreams in
the series m ig h t suggest th a t the m arriage was in d iffic u lty and found several:
(a) the stone from her engagement rin g is m issing; (b) her husband has
tuberculosis; (c) one o f her wom en friends is going through a divorce; and
(d ) a frie n d w ho is about to be m arried receives a lo t o f useless bric-a-brac
fo r wedding presents. I f the H a ll-V a n de Castle system had been available
w hen th is analysis was made, the case could have been im proved by com par
ing the dream er’s aggressions-per-character ra tio w ith her husband to the
same ra tio w ith o ther men. I f it had been higher w ith her husband th a n w ith
o ther men, and if the dreamer had had a low er rate o f frie n d ly in teractio ns as
w e ll, then the m etaphoric hypothesis w ould have been supported by means
o f a nonm etaphoric content analysis.
T w o later chapters in th is volum e provide methods th a t m ig h t aid in
the search for figurative m eaning in dream content. C hapter 4 suggests new
ways to conduct em pirical studies on m etaphors in dreams thro ugh the use
o f sophisticated software to search fo r phrases and strings o f words in large
numbers o f o n lin e dream reports. C hapter 5 presents findings w ith this
m ethod as one p a rt o f a study o f 3,116 dreams from one person over a 20-
year period. Because some o f the findings presented there co n tra d ict the
c o n tin u ity p rin c ip le , it may be th a t those findings in vo lve dream elements
th a t are figurative in nature. For exam ple, the series contains instances o f
TOWARD A NEUROCOGNIT1VE M ODEL OF DREAMS 35

the dream er rid in g horses or shooting guns w e ll, b u t contrary to expectations,
she does n o t ride or shoot in w aking life and is fe a rfu l o f both horses and guns.
T he p o ssib ility th a t some dreams may be based on fig u ra tive th in k in g
provides a way fo r a neurocognitive m odel to incorporate the in terestin g
idea th a t past experiences are sometimes used as personal m etaphors to
express current co n flicts th a t have sim ila r em otions and feelings at th e ir
core (K ram er et al., 1987). T h is idea comes from a study o f V ie tn a m veterans
w ho had recovered from th e ir posttraum atic stress disorder b u t returned to
the Veterans A d m in is tra tio n fo r help w hen w ar-related themes began to
appear in th e ir dreams in the face o f new life stressors, such as m a rita l
c o n flic t, conflicts w ith ch ild re n , or w ork-related tensions. In effect, the new
w ar-related dreams may have been conceptual blends th a t com bined past
experiences w ith aspects o f the stressful situations the veterans were endur
ing. T he resemblance is in the s im ila rity o f the feelings in b o th the war
and the new situ a tio n . “ I t ’s a war zone out there,” they m ig h t be th in k in g
in re la tio n to th e ir cu rre n t problems.
I f dream ing is in p a rt figurative, especially w ith regard to prim ary
m etaphors, resemblance m etaphors, m etonym ies, and conceptual blends,
th e n a neu ro cognitive m odel could advance in p a ra lle l w ith new understand
ings in cogn itive lingu istics. However, it s till w ould be necessary to do
the same kinds o f th e m a tic and H a ll-V a n de Castle co n te n t analyses to
understand any given series o f dreams, because m any resemblance m etaphors
and m ost conceptual blends are lik e ly to be unique to the dreamer. In
a d d itio n , to the degree th a t dreams are lik e proverbs and parables, it rem ains
necessary to study m any dreams in searching fo r the “ generic” o r u n d e rly
in g pattern .
For now , it needs to be stressed th a t there is little or no system atic
evidence th a t dreams make use o f the vast system o f fig u ra tive th o u g h t
available to m ost people in w aking life th ro u g h a co m b in a tio n o f develop
m ental experiences and c u ltu ra l heritage. O f a ll the possible connections
am ong the three areas o f dream research suggested in this chapter, the idea
th a t studies o f dream co n te n t may provide bridges to w aking figurative
th o u g h t is by far the m ost speculative. I t is also an issue th a t divides dream
theorists. For exam ple, b o th Foulkes (1999, p. 110) and H obson e t al.
(2000b), w ho disagree on m any issues, are together in doubting th a t unusual
constructions in dreams are m eaningful. For Foulkes, dreams reveal the
lim ite d nature o f the co g n itive a b ilitie s possessed by the sleeping brain ; fo r
H obson e t al., dreams are o fte n reflections o f the unique neurom odulation
o f the neural netw o rk fo r dream ing, a form o f d e liriu m durin g sleep.
Even if it turns out th a t dreams make little o r no use o f figurative
th o u g h t, a co g n itive theory is useful in e xp la in in g w hy dreams h o ld great
fascina tion fo r m any people in many d iffe re n t cultures: Dreams seem to have
parallels w ith w aking figurative thought. T he parallels w ith the m etaphoric

dim ensions o f w a king thought m ay be w hy some societies have used dreams
in th e ir cu ltu ra l practices and ritu a ls. In th a t sense, dreams have “ em ergent”
uses th a t have been developed in th e course o f h isto ry and passed on thro ugh
culture. T h is vie w also explains th e use o f dreams in psychotherapy: Dream
interpreters use m etaphoric in terpretations th a t are plausible to the c lie n t.
I t may be th a t dreams sim ply provide a platform from w h ich the c lie n t and
therapist, through a process o f n e g o tia tio n about m etaphoric meanings, can
develop a new n arrative about the c lie n t’s life . For the foreseeable future,
th e n , m etaphoric interpretations are the fo o l’s gold o f dream theories. W ith
th e ir g litte r o f seeming insight and the accom panying feelings o f enrich m e nt
and closure, m etaphoric in terpretations deceive interpreters and dreamers
alike.
T he neurocognitive m odel described in th is volum e also can in corp o-
rate the unexpected finding th a t nightm ares o fte n can be elim ina ted by
ha vin g people w rite out and visu a lly rehearse a new ending o f th e ir ow n
choosing fo r the dream (Krakow , K e lln er, Pathak, & Lam bert, 1995). T h is
process may be an instance o f th e cogn itive distancing th a t m any people
achieve by w ritin g about personal feelings and events (Pennebaker &
G raybeal, 2001; Pennebaker & K e o u g h , 1999; Pennebaker & Seagal, 1999).
In closing th is discussion o f dream ing and c o g n itio n , it is w o rth m en
tio n in g th a t a new neurocognitive m odel m ig h t tu rn o u t to be useful in
understanding th e developm ent o f consciousness. Foulkes (1990a, 1999)
offered th is fresh idea in lig h t o f his cross-sectional study o f ch ild re n ages
5 to 8. I f it is assumed th a t dream ing is the form th a t consciousness takes
d u rin g sleep, th e n th e origins o f consciousness can be explored by conducting
detailed studies o f the developm ent o f the a b ility to dream. As one part o f
th is general idea, Foulkes fu rth e r suggested th a t the a b ility to include oneself
in a dream, w h ich is n o t fu lly developed u n til around age 8, may be an
in d e x o f when a c h ild has a fu ll sense o f self. Fie reached this conclusion
after fin d in g th a t three waking tests designed to assess the developm ent o f
the “ s e lf’ concept d id n o t correlate w ith each o ther and did n o t predict
the in clu sio n o f the dreamer as a character in his or her dream reports
(Foulkes, 1999, p. 95).
C O N C L U S IO N
O th e r possible lin ks may e xist among the three areas o f dream research
discussed in this chapter. H ow ever, enough has been said to dem onstrate
th a t there is a large body o f established em pirical findings upon w h ich to
base a new m odel. M oreover, the research tools, such as the rapid advances
in neuroim aging and neurochem istry, are now available to do th e many
studies th a t w ould be necessary to test and develop the model. A s noted

earlier, th e grow ing num ber o f neuropsychologists in c lin ic a l settings may
be an im p o rta n t resource fo r developing th is m odel, because th e y could
easily screen fo r changes in dream ing as they exam ine people w ith lesions
in re le va n t areas o f the brain.
T he advent o f personal com puters and the constant im provem ents in
softw are are also im p o rta n t because they have made co n te n t analysis some-
w h at less labor intensive and far more accurate th a n it was in the past.
These advances include a spreadsheet th a t calculates a ll the H a ll-V a n de
C astle co n te n t indicators (Schneider &. D om hoff, 1995) and is discussed
in chapter 3. In add ition , a new search program allows users to fin d single
words, strings o f words, or phrases in the m ore th a n 11,000 dream reports
available on D ream Bank.net (Schneider & D om hoff, 1999). T h is search
program m ay prove especially useful fo r conducting th e m eta phoric studies
th a t are necessary to determ ine the degree to w h ich there are lim its to the
m eaning in dreams. T he use o f D ream Bank.net fo r new approaches to
co n te n t analysis is dem onstrated in chapter 4.

2
METHODOLOGICAL ISSUES IN
THE STUDY OF DREAM CONTENT
T his chapter examines the m ajor m ethodological issues th a t arise in

the k in d o f scientific studies o f dream content th a t w ould be necessary for
the fu ll elaboration o f the neurocognitive m odel presented in chapter 1.
Those issues include the degree to w hich dream reports actually reflect the
dream ing experience, the usefulness o f various methods o f colle ctin g dream
reports, the representativeness o f the people who provide dreams, the value
o f several d iffe re nt methods o f dream analysis, and problems having to do
w ith statistics and sample size. T his chapter suggests th a t although many
dream studies have used inadequate methods o f dream co lle ctio n and data
analysis, sound methods exist fo r co lle ctin g representative samples o f dream
reports as w ell as analyzing dream content. These conclusions provide the
basis fo r a detailed presentation o f the H a ll-V a n de Castle coding system
in chapter 3, and fo r the kinds o f studies presented in chapters 4 and 5 th a t
could be used in developing the m odel.
The study o f dreams is a unique to p ic fo r psychological investigation
fo r several reasons. First, as shown thro u gh the experim ental studies discussed
in chapter 1, it has n o t been possible to shape dreams to any appreciable
e xte n t by the ap plicatio n o f external s tim u li o r the use o f verbal instructions,
a s itu a tio n th a t makes the experim ental m ethod less useful th a n it usually
is. Second, dreams cannot be observed by anyone b u t the dreamer w hile
they are happening, so the observational m ethods so im p o rta n t to some
realms o f scie n tific endeavor are com pletely irre le va n t. T h ird , dreams cannot
be reported by dreamers w h ile they are dream ing; it is therefore d iffic u lt to
use the m ethods fo r studying subjective experiences under w aking co n ditio n s,
except after im m ediate awakenings (Fiss, 1983, 1991).
For a ll o f these reasons, the dream experience is usually available to
investigators o n ly as a verbal o r w ritte n re p ort o f a w aking m em ory. T h e
fle e tin g rare exceptions are m entioned in the n e x t section. Thus, th e study
o f dream c o n te n t is generally tw o steps rem oved fro m the process o f dream ing.
T h is p o in t in e v ita b ly raises questions concerning th e degree to w h ic h dream
reports fa ith fu lly represent the process o f dream ing. Skeptics therefore some-
39
tim es cla im th a t the dreams th a t are recalled are bound to be those th a t
are p a rticu la rly salient, a critic is m th a t im plies th a t samples are selective
and unrepresentative. However, despite a ll these p o te n tia l problem s, e v i
dence has in dica ted th a t dream reports provide a sound basis fo r understand
ing b o th the form al structure and co n te n t o f dream ing, as discussed in the
n e xt section.
T H E A C C U R A C Y A N D R EPR ESENTATIVEN ESS

O F D R E A M REPORTS
O ne o f the m ost curious and least understood facts about dreams is

how few o f them are recalled in everyday life com pared w ith the recall rate
o f 80% to 90% fo r norm ative samples o f young adults w hen awakened in
the sleep laboratory. T he riv a l neurophysiological and co g n itive explanations
fo r th is la ck o f re ca ll are discussed b rie fly in chapter 6. For purposes o f this
chapter, how ever, there is a m ore im m ediate and tractable question: A re
the few dreams th a t people re ca ll a representative sample o f th e ir dream life?
Because o f the inaccessibility o f dream ing to outside observers, any
argum ent fo r the accuracy o f dream reports m ust begin w ith evidence th a t
a phenom enon exists (i.e., dream ing) th a t the dream report describes. I t
th e n has to dem onstrate th a t at least some dream reports— th a t is, those
collected in sleep laboratories from n ig h t awakenings— are accurate reflec
tio n s o f dream ing. N e x t, it needs to show th a t the dream reports th a t seem
m ost questionable— those w ritte n dow n at hom e or collected w ith a standard
fo rm in a group setting— are enough lik e laboratory reports to be regarded
as useful.
T h e fa ct o f dream ing w ould seem to be w e ll established by findings
in sleep laboratories since the early 1950s, b u t skeptics can n o te th a t even
after a ll o f those studies, the self-report evidence is consistent w ith the
p o ssib ility th a t dream ing happens during the process o f aw akening or o f
te llin g the dream. Eye m ovem ents and a unique b ra in wave p a tte rn are n o t
the same th in g as dream ing. U ltim a te ly , no d e fin itiv e way may establish
the fact o f dream ing beyond the u b iq u ity o f th is subjective experience, but
tw o unusual phenom ena go a good part o f the way in dealing w ith this
issue: sleep ta lk in g and REM sleep-behavior disorder. C areful studies o f a
large num ber o f sleep-talking episodes show th a t those th a t happen in REM
are consistent w ith w h a t the dreamer reports after an im m ediate awakening
(A rk in , 1981). T h is fin d in g provides objective evidence corroborating w hat
the sleep ta lke r th e n reports. M ore dram atically, the actions th a t occur
durin g episodes o f R E M sleep-behavior disorder are o fte n found to fit the
actions th a t patients report w hen they are awakened (M ahow ald & Schenck,
2000). These rare exceptions to the inaccessibility o f dream ing to th ird -

person observation provide independent evidence th a t dream reports can be
accurate, at least w hen the reports occur shortly after the dream experience.
Even if the fa ct o f dream ing is granted, it remains the case th a t self-
report data are notoriously unreliable o n everything from v o tin g to church
attendance to eating habits. They are o fte n based on im pression management
and c u ltu ra l stereotypes. I t is therefore im p o rta n t to show th a t dream reports
are n o t typ ica l self-report data. People do n o t feel responsible fo r th e ir
dreams, so they have no reason to w ith h o ld in fo rm a tio n or d isto rt th e ir
reports (Foulkes, 1979; H a ll & V a n de Castle, 1966). They tend to experi
ence dreams as som ething th a t happens to them , n o t as som ething fo r
w h ich they are personally responsible. M any people are therefore w illin g
to volunteer th e ir dreams if asked. T he w illingness o f people to report
accurately is increased w hen th e ir reports are anonymous, as they are for
m ost o f the group studies cited in th is book. In add ition , people have no
reason to d isto rt dreams w ritte n in a personal dream jo u rn a l n o t o rig in a lly
m eant fo r la te r researchers.
T he atypical nature o f dream reports w hen compared to o ther types
o f self-reports is dem onstrated em p irica lly by com paring them w ith w hat
people claim they dream about in response to questionnaires, where they
provide c u ltu ra lly stereotypic distortion s. Four d iffe re n t samples o f college
students were asked about the frequency w ith w h ich they dreamed about
aggressive, frie n d ly, and sexual interactions; they th e n w rote dow n five
dreams over a 2-week period. In response to the questionnaire, they claim ed
th a t they dream m ost frequently o f frie n d ly interactions, th e n sexual interac
tions, follow ed by aggressive interactions. The actual order in the dreams
they la te r turned in to the investigators and in a ll o ther carefully analyzed
samples was (1 ) aggressive in teractio ns, (2) frie n d ly interactions, and
(3) sexual in teractio ns (B ernstein &. B e licki, 1995). Thus, the rank-order
corre latio n between expressed opinions and a co n te n t analysis o f social
interactions is negative. These findings show th a t self-reports on dream
co n te n t are worthless at best and o fte n m isleading. Even more im portant,
they suggest th a t a descriptive report o f a dream experience is n o t a typ ica l
form o f self-report in fo rm a tio n : T he dream experience is so dire ct, im m edi
ate, and com pelling th a t a report on it is n o t lik e ly to be distorted by filte rin g
it through cu ltu ra l beliefs. T his p o in t is especially strong fo r awakenings in
the sleep laboratory w h ile the person is in the process o f dream ing.
As noted b rie fly in chapter 1, several d iffe re n t studies o f dream co n te n t
in the laboratory provide a consistent picture o f dream con te n t as being
m ore mundane and coherent th a n standard cu ltu ra l stereotypes, although
dreams have some unusual aspects as w e ll (C avallero &. Foulkes, 1993;
Foulkes, 1985; M eier, 1993; Snyder, 1970; Strauch &. M eier, 1996). O th e r
laboratory studies show th a t dream co n te n t does n o t d iffe r from early to
late REM periods (D em ent &. W o lp e rt, 1958a; D o m h o ff & Kam iya, 1964b;
Foulkes, 1966; H a ll, 1966b; Strauch & M eier, 1996; Trosm an, R echtschaf-
fen, O ffenkrantz, & W o lp e rt, 1960), even though REM periods become
longer and more intense thro ugho ut the n ig h t (A n tro b u s e t al., 1995). O ne
study o f five particip ants d id fin d one difference am ong the m any comparisons
it made between early and la te r REM periods: a greater num ber o f references
to the past in la te r R E M periods (Verdone, 1965). H ow ever, th a t fin d in g
was n o t replicated in a larger study w ith 11 particip ants over a greater
num ber o f nights (H a ll, 1966b). T he findings o n the consistency o f dream
co n te n t throughout the n ig h t are im p o rta n t because they show th a t any
bias tow ard la te -n ig h t dreams in a sample o f dream reports does n o t produce
a biased sample o f dream content.
Studies in the laboratory com paring dreams reported after n ig h ttim e
awakenings w ith those th a t are s till remembered in the m orning provide a
way o f assessing the saliency issue in dream samples. Such studies are n o t
a com plete sim u la tio n o f dream recall at hom e because a “p rim in g ” has
occurred as a result o f the n ig h t reports, but they provide a close approxim a
tio n (Baekland & Lasky, 1968; M eier, Ruef, Zeigler, & H a ll, 1968; Strauch,
1969; T rin d e r & Kram er, 1971). These studies firs t show th a t m uch dream
re ca ll comes from dreams late in a sleep period, as has also been found in
studies o f participants asked to keep dream diaries at hom e or leave reports
on a telephone answering m achine (B e lic k i, 1987; D om hoff, 1969). T h is
fin d in g is im p o rta n t because it shows th a t recency, n o t saliency, is the m ost
im p o rta n t facto r in dream recall. M oreover, these studies also dem onstrate
th a t the d u ra tio n o f a dream, as indexed by the le n g th o f the report given
at the n ig h t awakening, is an im p o rta n t pre d icto r o f m orning dream recall.
F in a lly, these studies provide evidence th a t the em otion al in te n s ity o f the
dream co n te n t also can affect recall.
In the m ost sustained o f these studies, M eier et al. (1968) compared
the n ig h t and m orning recall o f a m an who was awakened at the end o f
every REM period fo r 45 nights. O f the 138 dreams he reported w hen
awakened, he recalled 88 in the m orning. Dreams from the last 105 m inutes
o f the sleep period were recalled 83% o f the tim e in the m orning, compared
w ith 63% , 55% , and 54% fo r th e three preceding 105-m inute periods. Long
dream reports were recalled 87% o f the tim e, com pared w ith 48% recall
fo r short reports. Dream reports judged as m ore em o tio n a lly intense by the
dream er and an independent rate r were recalled 83% o f the tim e, compared
w ith 56% fo r those judged less intense. A n analysis o f the interactions
among these three variables found th a t recency can compensate fo r b o th
shortness o f reports and low in te n sity and th a t le n g th can compensate fo r
low in te n sity. These results strongly suggest th a t recency and d u ra tio n lead
to m any everyday dream reports th a t are n o t unusually h ig h in saliency.
Because it already has been shown th a t dream co n te n t does n o t change
thro ugho ut the n ig h t, it therefore follow s th a t recency and d u ra tio n are

p ro vid in g a more representative sample o f dreams th a n an exclusive emphasis
on saliency w ould suggest.
T he findings presented in the previous paragraphs establish th a t dream
reports collected in the sleep laboratory provide a reasonably representative
sample o f dream life . W ith in th is context, studies th a t d ire ctly compare
dreams reported after laboratory awakenings w ith dreams w ritte n down
at hom e by the same participants provide a good basis fo r judging the
representativeness o f everyday dream recall. A lth o u g h one early study
claim ed there were some differences between w hat came to be called “ lab”
and “ hom e” dreams (D o m h o ff & Kam iya, 1964a), p rim a rily on indicators
o f aggression, five later studies in trodu cing controls fo r methods o f reporting
and o ther confounds suggested no im p o rta n t differences (Foulkes, 1979;
H e yn ick & dejong, 1985; Strauch St M eier, 1996; W eisz & Foulkes, 1970;
Zepelin, 1972).
T he most com prehensive com parison o f dream con te n t inside and
outside the sleep laboratory was carried out by H a ll (1966a) as one p a rt o f
a study concerning the many p o te n tia l problem s th a t m ig h t affect the
accuracy and representativeness o f dream reports. T h e results from this study
are w o rth recounting in d e ta il because they make the best possible case
th a t everyday dream recall can provide a good sample o f dream life . T he
study was conducted in a large house in a q uiet residential neighborhood,
where the 11 young adult male participants between ages 19 and 25 could
have th e ir ow n sleeping quarters fo r a m o n th and report dreams in the least
thre ate nin g atmosphere possible, thereby m in im izin g any in h ib ito ry effects
fro m the usual laboratory setting in a science or m edical b u ild in g . A to ta l
o f 414 planned awakenings to o k place.
Because the aim o f the study was to determ ine the conditions th a t led
to the m ost representative sample o f dream reports, seven adjustm ent nights
were provided before the form al co lle c tio n o f reports began. T he tim e it
to o k the participants to fa ll asleep was noted. In a carefully balanced design,
o n ly one awakening <occurred on some nights, and m u ltip le awakenings
occurred on others. Dreams were collected by means o f tape-recorded reports
fro m the first four REM periods o f the n ig h t under b o th the single-aw akening
and the m ultiple-aw akening schedules. P articipants were asked after each
awakening to estim ate how long they had been dream ing. They also were
asked to rate the c la rity o f th e ir recall, the vividness o f the dream, the
em otion al in te n sity o f the dream, and w hether the dream to o k place in the
past or present. U nplanned “ spontaneous” awakenings were noted, and any
dream reports from them were transcribed fo r com parison w ith dream reports
from scheduled awakenings. F inally, participants w rote dow n any dreams
they remembered at hom e fo r a 2-week period. Some w rote th e ir dream
reports before th e ir stay in the laboratory, some before and during th e ir
stay, and some during and after th e ir stay. T he goal was to have at least 15
M ETH O D O LO G IC AL ISSUES 43
dreams w ritte n dow n a t hom e by each person, but one p a rtic ip a n t w rote
dow n o n ly 11 and another d id n o t w rite down any.
P articipants had little d iffic u lty adjusting to the laboratory situatio n.
B eginning w ith the first n ig h t th a t the EEG m achine was turned on, w h ich
was the th ird adjustm ent n ig h t, it to o k them no longer to fa ll asleep than
it d id on la te r nights (H a ll, 1966b, p. 38). T hen, too, the sm all percentage
o f dream reports th a t included allusions to the experim ental s itu a tio n (7.2%
to 13.5% ) d id n o t vary from the fo u rth adjustm ent n ig h t, w hen they were
firs t awakened to report a dream, to the end o f th e ir laboratory visits (H a ll,
1966b, p. 32). T h is range is m uch low er and narrow er th a n the 20% to 30%
reported in previous studies (D em ent, K ahn, & Roffwarg, 1965; D o m h o ff &
Kam iya, 1964a; W h itm a n , Pierce, Maas, & Baldridge, 1962). C o n tra ry to
expectations, w h ic h were based on a study by D em ent and K le itm a n (1957b)
th a t reported th a t five participants could corre ctly distinguish between awak
enings after 5 o r 15 m inutes o f R EM dream ing, no co rre la tio n was found
between the am ount o f R EM tim e before an awakening and p a rticip a n ts’
estimates o f how long they had been dream ing (H a ll, 1966b, pp. 10-11,
3 8 -3 9 ).
T he m ost im p o rta n t result from the comparisons o f dream reports
collected in d iffe re n t ways w ith in the laboratory was th e ir general sim ila rity ,
w hether from single o r m u ltip le awakenings, or early o r late R E M periods,
on the several H a ll-V a n de Castle categories th a t were used. M oreover,
the results fo r the ra tin g scales responded to by the dreamers at the tim e
o f aw akening were sim ila r in show ing no differences, except th a t particip ants
reported b e tte r re ca ll and greater c la rity fo r each successive aw akening on
nights w hen there were m u ltip le awakenings, a fin d in g th a t could be the
result o f a practice effect. In ad d itio n , the 57 dream reports from particip ants
who had spontaneous awakenings did n o t d iffe r from those collected on
th e ir single-aw akening nights (H a ll, 1966b, pp. 2 5 -2 6 ). These results re p li
cate and extend studies show ing th a t it is possible to co lle ct a representative
sample o f a person’s dream life in the sleep laboratory.
Several differences were found betw een laboratory and hom e dream
reports on 26 com parisons fo r each p a rtic ip a n t using the nonparam etric
W ilc o x o n m atched pairs, signed-rank test. M ost o f the sta tistica lly sig n ifica n t
differences concerned the larger num ber o f aggressions and m isfortunes in
hom e dream reports. A t the same tim e, hom e and laboratory dreams had
few differences in types o f characters and no differences in the percentage
o f dreams w ith at least one “ bizarre” (i.e ., unusual) elem ent. T h e fin d in g
on bizarre elem ents does n o t support a fin d in g o f more such elem ents in
laboratory dream reports in the study by D o m ho ff and Kam iya (1964a). A
later study by H u n t, O g ilvie , B e licki, B e licki, and A ta lic k (1982) also
reported no differences in bizarreness between hom e and laboratory
dream reports.

T he nature o f the differences between dream reports under the tw o
cond ition s was summarized by findings w ith w hat H a ll term ed the dramatic
intensity index, w h ich he calculated by adding together a ll aggressions,
frie n d ly interactions, sexual a ctivitie s, successes, failures, good fortunes,
and m isfortunes. T h is index showed consistent differences between reports
w ritte n at hom e and laboratory dreams collected through e ith e r single or
m u ltip le awakenings fo r b o th early and la te REM periods (H a ll, 1966b).
T h is fin d in g provides support fo r the cla im th a t there is a saliency bias in
hom e dream samples. However, a reanalysis o f the o rig in a l codings revealed
th a t m ost o f the variance in this index was provided by aggressive elements
and th a t the overall picture is very d iffe re n t w hen effect sizes are considered
(D o m h o ff & Schneider, 1999).
T he reanalysis o f the H a ll and V a n de Castle data was based on eight
young m en w ho provided at least 15 hom e dreams and at least 34 laboratory
dreams. W hen m ore th a n 15 hom e dream reports were provided, the first
15 were used. W hen more than 34 laboratory dream reports were available,
reports from adjustm ent nights and spontaneous awakenings were elim inated
first, and then an equal num ber o f single-aw akening and m ultiple-aw akening
reports were rem oved from the sample if a d d itio n a l reports had to be dis
carded. T he result was a group sample o f 120 dream reports w ritte n down
at hom e and 272 dream reports transcribed from tape-recorded reports in
the laboratory.
Three o f the four sta tistica lly sig n ifica n t differences found in the reanal
ysis in vo lve d aggression. T he hom e dream reports contained a higher per
centage o f dreams w ith at least one in c id e n t o f aggression; a h igher rate o f
aggressions per character; and a h ig h e r percentage o f aggressions th a t were
physical in nature, as defined by destruction o f personal property, chases,
physical attacks, and murders. O n the basis o f an effect-size statistic explained
in chapter 3, the effect sizes were sm all even w hen the differences were
sta tistica lly sig nifica nt, except in the case o f the physical aggression percent
age, w h ich showed a large difference. Because the num ber o f dreams in volve d
in th is analysis is large, especially compared w ith m ost published dream
studies, it is lik e ly th a t the findings on effect size are accurate.
T he higher frequency o f aggression in hom e dream reports supports
the concern th a t there is some selective recall in everyday dream reports.
Even here, how ever, it is notew orthy th a t 44% o f the dreams did n o t contain
any form o f aggression, w hether physical o r nonphysical, and 72% were
w ith o u t any physical aggression. These findings are fa irly sim ila r to those
obtained w ith a norm ative sample o f young m en (see chapter 3 ), in w h ich
53% o f the reports had no form o f aggression and 74% had no physical
aggression. M oreover, only 10% o f the hom e and laboratory reports in the
H a ll study had a bizarre elem ent, w h ich is low er than w hat m ig h t be expected
if o n ly atypical con te n t is recalled. I f dream reports have a strong bias fo r
atypica l elements, th e n it also m ig h t be predicted th a t a large num ber o f
sexual dreams w ould be reported, b u t o n ly 9% o f the hom e reports in the
H a ll study contained so m uch as a sensual hug o r kiss, com pared w ith 5%
o f the laboratory reports. A reference to sexual intercourse occurred in o n ly
2.5% o f the hom e reports and 0.7% o f th e laboratory reports; the form er
num ber is n o t th a t d iffe re n t from th e figure o f 3.4% in the H a ll-V a n de
C astle male norm ative sample described in chapter 3.
T h e fact th a t the m ost consistent differences between laboratory and
hom e dream reports relate to aggression fits w ith findings on the v a ria b ility
o f aggression in c o n ju n c tio n w ith several o th e r factors. First, the difference
in aggression between th e early teens and young adulthood is the largest
difference between the tw o age groups (A v ila -W h ite , Schneider, &
D om hoff, 1999), Second, a decline in aggression may occur in o ld age
(H a ll & D om hoff, 1963b; Zepelin, 1980), although some results from
lo n g itu d in a l studies make th is cross-sectional fin d in g less ce rta in (D om hoff,
1996, chapter 7). T h ird , large variations exist from culture to cu ltu re in
aggression (D om hoff, 1996, chapter 6). F ourth, large in d iv id u a l and gender
differences have been found on some measures o f aggression in dreams
(D o m h o ff, 1996, chapter 8; H a ll & D om hoff, 1963b; P aolino, 1964). In
short, these findings suggest th a t variations in aggression could be valuable
in developing a neurocognitive m odel because aggression m ig h t be especially
sensitive to drug effects or lesions in one or another p a rt o f the neural
n e tw o rk fo r dream ing.
In lig h t o f the several d iffe re n t arguments and laboratory studies pre
sented in this section, it seems safe to conclude th a t the dreams people
recall in the laboratory and at hom e are, in fact, a reasonably representative
sample o f dream life . I f reports obtained im m ediately after awakenings in
the sleep laboratory are taken as an excelle nt sta rtin g p o in t fo r assessing
the adequacy o f hom e dream reports, th e n the saliency cla im is lim ite d to
the tendency fo r greater aggression in hom e dream reports.
M E T H O D S FO R C O L L E C T IN G D R E A M REPORTS
T h e arguments and evidence presented in the previous section do

n o t m ean th a t a ll m ethods fo r co lle ctin g dreams are equally useful. Strengths
and weaknesses are connected w ith each o f the five m ain m ethods. T hree
are generally quite useful: awakenings in the sleep laboratory; anonym ous
w ritte n reports collected in group settings; and personal dream journals
kept over the space o f m onths, years, o r decades. T he rem aining tw o—
dream diaries kept at hom e fo r a week or tw o at the request o f investigators
and dreams collected in the psychotherapy re la tio n sh ip — have serious
weaknesses.

Sleep Laboratory Awakenings
Sleep laboratories are the best source o f dream reports because they
provide the op p o rtu n ity fo r co lle ctin g large representative samples o f people’s
dreams under co ntrolled cond ition s (Foulkes, 1966, 1985). Aw akenings
during REM or from Stage II N R E M late in the sleep period m axim ize the
p ro b a b ility o f recall and make it possible to co lle ct as m any as fo u r or five
dreams in a single n ig h t (A ntro bus e t ah, 1995; Fosse, S tickgold, & Hobson,
2001; Foulkes, 1979). However, there are some problem s w ith th is m ethod.
T he m ain problem is th a t it is a costly and tim e-consum ing process. Labora
tories are expensive to equip and staff, and participants have to be paid.
T he sleep laboratory has been especially d iffic u lt to use in the U n ite d States
since the 1970s because o f the decline in outside funding fo r dream research
(Foulkes, 1996a).
Even though several dreams can be collected each n ig h t, it s till can
take m any m onths to obtain 10 or m ore dreams fro m each o f a dozen
participants. In addition, frequent awakenings can be ta xin g fo r participants,
who o fte n resist fu ll awakenings and com plete reporting. M oreover, staying
up m ost o f the n ig h t several tim es a week can be onerous for investigators.
M ost Recent Dream M ethod
T he m ost objective and structured co n te xt fo r the e fficie n t and in e x

pensive co lle c tio n o f dream reports outside the sleep laboratory is a group
setting in w h ich reports can be w ritte n on a standardized form by anonymous
participants, who reveal o n ly basic background in fo rm a tio n such as age and
gender. Pioneered by H artm ann, E lkin , and Garg (1991), the “ m ost recent
dream” m ethod provides a way to co lle ct dreams from people in classrooms
and w a itin g rooms in m any d iffe re n t regions or countries in the space o f 15
to 20 m inutes fo r adults and 20 to 30 m inutes fo r teenagers and preadolescent
c h ild re n (A v ila -W h ite et al., 1999; D om hoff, 1996).
T he m ain drawback o f th is m ethod is th a t the available tim e does n o t
usually p e rm it co lle c tio n o f any personality or co g n itive measures on the
people p ro vid in g the reports. In studying children, serious problem s rem ain
w ith determ ining the age at w h ich researchers can be confident in the
a u th e n tic ity o f the dream reports. W h e n the frequencies fo r “ creatures,”
video game characters, and physical aggressions in these reports are compared
w ith findings from laboratory reports from ch ild re n o f the same age, there
is reason to believe th a t young ch ild re n up to at least age 10 are using th e ir
w aking im aginations to provide a report th a t fits cu ltu ra l stereotypes about
the nature o f dreams. T h is problem may be especially great w ith boys. A t
th is p o in t, the m ethod cannot be recommended fo r use w ith ch ild re n below
the six th grade (S aline, 1999).
METHODOLOGICAL ISSUES 47
T he standardized H a ll-V a n de C astle form on w h ich the m ost recent
dreams are collected begins by sim ply asking the respondents to “ w rite
dow n the m ost recent dream” they can remember, “ w hether it was last
n ig h t, last week, o r last m o n th ” (D om hoff, 1996, p. 67). T o reinforce the
emphasis on the last dream recalled and to make it possible to e lim ina te
dreams fro m m onths or years in th e past if the researchers so desire,
pa rticip a n ts are also asked to w rite dow n the date on w h ich they th in k
the dream occurred. T he in struction s th e n exp la in w hat the report should
include, using language developed by H a ll (1951, 1953c) fo r c o lle ctin g
dreams from college students.
T he usefulness o f the m ethod w ith adequate sample sizes has been
dem onstrated m ost d ire ctly by com paring the results fro m A v ila -W h ite ,
Schneider, and DomhofPs (1999) m ost recent dream study o f 12- to 13-
year-old girls and boys w ith the findings fo r the same age group in tw o
lo n g itu d in a l studies in the sleep laboratory (Foulkes, 1982; Strauch & Leder-
bogen, 1999). W here d ire ct com parisons were possible, there were many
sim ila ritie s. T he m ethod also receives support because m ost recent dream
samples collected from college students m atch the findings from a large-
scale norm ative study discussed in the n e x t chapter (D om hoff, 1996).
Personal Dream Journals
Dream journals, w h ich are called “ dream series” in the research lite ra
ture, are an underused source o f dream reports, even though they are a form
o f “ personal docum ent” long recognized in psychology as h aving the p o te n tia l
to provide insights in to personality and co g n itive styles (A llp o rt, 1942;
B aldw in, 1942; S m ith, 2000). N o n re a ctive a rch iva l sources, such as dream
journals, have the advantage o f n o t being influenced by the purposes o f the
investigators w ho analyze them . G eneralizations th a t are based on nonreac-
tiv e a rch iva l data are considered m ost im pressive w hen they derive from a
d ive rsity o f archives lik e ly to have d iffe re n t types o f possible bias (W ebb,
C am pbell, Schwartz, Sechrest, & G rove, 1981). A s summarized in chapter
1, dream journals kept fo r d iffe re n t reasons lead to the fin d in g o f great
consistency in dream co n te n t over tim e
For a ll th e ir p o te n tia l usefulness, how ever, dream journals are n o t
w ith o u t th e ir drawbacks. Even after show ing in itia l w illingness, some people
may n o t w a nt to provide a ll o f th e ir dreams fo r scie n tific scrutiny. Journals
may have unexpected gaps or omissions. Care m ust be taken to ensure
th a t they are authen tic. Journals from dream popularizers w ho have made
enthusiastic claim s about th e ir dreams probably should be avoided, as should
dream journals posted o n the In te rn e t by in dividuals. Thus, personal dream
journals are best used selectively.

In future studies using dream journals, it m ig h t be w o rth w h ile to
co lle c t a laboratory sample o f dreams from the person as w e ll or to co lle ct
dreams at hom e thro ugh systematic awakenings using a sleep m o n ito rin g
device called “ the N ig h tca p ,” w h ich is a sm all portable u n it th a t detects
eyelid and head m ovem ents using separate m ovem ent sensors. T he sensors
can be attached to a headband or taped d ire c tly to the forehead. Signals
are sent to a sm all, portable bedside u n it w eighing less than 5 pounds.
T h e in fo rm a tio n it stores can be sent to a personal com puter and used
to perform awakenings in REM or N R E M w ith h ig h levels o f accuracy
(A jilo re , S tickgold, R ittenhouse, & Hobson, 1995; Stickgold, Pace-Schott,
&. Hobson, 1994).
T he use o f dream series fo r the study o f dreams was introduced in to
th e c lin ic a l research tra d itio n by W ilh e lm Stekel (1911) and H avelock E llis
(E llis, 1928). T he m ost in flu e n tia l early analysis o f a dream series was
published in 1935 by Jung to dem onstrate the search fo r personal in te g ra tio n
and wholeness in 74 dreams from a natural scientist suffering from depression
and d rin k in g problem s. T h is dream d ia rist was subsequently revealed to be
W olfgang Pauli, one o f the m ost b rillia n t physicists o f the 20th century and
the w inne r o f a N o b e l Prize in 1946 (L in d o rff, 1995; Zabriskie, 1995). Some
o f these dreams are available in books on Pauli (M eier, 1992; von M eyenn,
1993); many others are in P auli’s unedited correspondence w ith M arie-
Louise von Franz, w h ich is stored in the Pauli A rchives at the E T H -
B ib lio th e k in Zurich, Sw itzerland. A H a ll-V a n de Castle co n te n t analysis
o f th is unique series m ig h t be w o rth w h ile once it is possible to assemble a ll
the pieces in one place.
Jung’s o rig in a l analysis o f the Pauli series was expanded in 1944 to
include 400 dreams (Jung, 1974; L in d o rff, 1995). A lth o u g h the prim ary
emphasis is on how symbols fo r psychological wholeness relate to religious
themes and the practice o f alchem y in the M id d le Ages, a footnote at the
end contains the results o f a q u a n tita tive analysis in w h ich the 400 dreams
are divided in to e ight groups o f 50. I t claim s th a t the occurrence o f the
“m andala m o tif’ (i.e ., the sym bolic expression o f the search fo r wholeness)
increases from a range o f 2 to 9 per 50 dreams in the first fo u r sets to 11
to 17 in the second four sets (Jung, 1974, p. 296).
Dream series kept fo r personal, a rtistic, or in te lle c tu a l reasons have
been the basis o f the most systematic studies. For example, S m ith and H a ll
(1964) used a jo u rn a l kept on and o ff fo r 50 years fo r personal interest to
show th a t the dreamer d id n o t dream m ore o f the past w hen she was in
her 70s than she d id w hen she was under 40. H e r jo u rn a l also showed a
consistency over m any decades in several themes th a t are spelled out later
in this chapter. T he dreams are now available on http://w w w .D ream B ank.net
under the pseudonym “D orothea.”
Chapters 4 and 5 provide new evidence fo r the p o te n tia l o f th is source
o f useful data w hen it is com bined w ith rigorous m ethods o f co n te n t analysis.
B rie f Dream D iaries
T he m ost freq uently used m ethod o f c o lle ctin g dreams outside the
laboratory is to ask participants, usually h ig h school or college students, to
keep a dream diary fo r a period o f 1 or 2 weeks or u n til they have w ritte n
dow n a prescribed num ber o f dreams. T he m ethod has the advantage o f
being easy and inexpensive. I t has led to some useful collectio ns o f dream
reports from g ifte d girls ages 8 to 13 (L a tta , 1998), conscientious volu n te e r
college students (H a ll, 1947, 1953c; T onay, 1990/1991), and b lin d m en
and w om en (H u ro vitz et al., 1999), m ost o f w h ich are now available on
Dream Bank.net. I t reaches its highest and m ost useful le vel w hen the diary
is based on dreams collected during the n ig h t w ith the aid o f th e N ightcap
fo r m o n ito rin g sleep at hom e (Fosse e t al., 2001; S tickg old et al., 1994).
W h e n used w ith o u t the aid o f the N ightcap, th is m ethod has m any
drawbacks and o fte n leads to h ig h ly inadequate samples. F irst, it can take
weeks or m onths to o b ta in even fo u r dreams, as seen in a study o f gender
differences by B ursik (1998), in w h ich it to o k 4 m onths to o b ta in four
dreams from 40 m en and 40 wom en in one o f her undergraduate psychology
classes, and in studies by Lortie-Lussier and co-workers, in w h ic h it usually
to o k several weeks to o b ta in a m inim u m o f tw o dreams fro m the adult
wom en and m en in th e ir studies (C ote et al., 1996; Lortie-Lussier, Schwab,
& de K o n in ck, 1985).
Second, a large m in o rity o f participants drop out or tu rn in o n ly one
or tw o dreams, leading to questions about the representativeness o f those
who do tu rn in dreams. For example, b o th B uckley (1970) and H ow ard
(1978) reported d iffic u ltie s in obta in in g com pleted dream diaries fro m th e ir
teenage volunteers, especially from boys. In a study o f teenagers’ dreams,
W inegar and L e v in (1997) had o n ly 182 in itia l volunteers o u t o f the 550
students in the classrooms they visited. T hen, o n ly 115 o f those 182 turned
in at least tw o dreams o f 35 words or m ore, w h ich is a m inim u m report
le ngth. By contrast, dream reports collected fro m teenagers in laboratories
and classrooms are usually between 100 and 300 words in le n g th (A v ila -
W h ite et al., 1999; Foulkes, 1982; S trauch & Lederbogen, 1999).
Samples o fte n consist o f d iffe re n t numbers o f reports from each p a rtic i
pant, a situ a tio n th a t raises questions about how to standardize the c o n trib u
tio n o f each p a rtic ip a n t to the to ta l sample. S olutions to th is problem th a t
draw o n ly one or tw o dreams from each dream diary unnecessarily waste
data and lead to such sm all sample sizes th a t any differences th a t m ig h t
exist cannot be detected. F in a lly, the demand characteristics o f a dream -
diary study can be extrem ely strong, especially w hen researchers have to

prod unm otivated participants one or more tim es to w rite dow n th e ir dreams.'
Such pressures increase the p ro b a b ility o f hasty or confabulated reports. In
a survey o f several hundred students in a psychology class at the U n ive rsity
o f C a lifo rn ia , Berkeley, most o f whom had chosen n o t to v o lu n ta rily keep
a dream diary fo r 2 weeks, 43% said th a t they w ould be lik e ly to make up
dreams if required to tu rn in dream reports as p a rt o f a course assignment
(Tonay, 1990/1991). T h is is one o f several problem s in a flawed study
o f gender differences in dreams by Bursik (1998), w h ich is critiq u e d by
D o m h o ff (1999b).
The Psychotherapy Setting
W o rk w ith clients by Freud, Jung, and th e ir m any co-workers was a

m ajor source o f dream reports fo r the analysis o f dreams in the first h a lf
o f the 20th century. Such reports have the virtu e o f ric h accom panying
biographical and fantasy m aterials, and they can be analyzed in great d etail
as p a rt o f the person’s psychotherapy. For a ll th e ir usefulness in the past as
a source o f hypotheses concerning dream m eaning, the nature o f the thera
peutic relatio nship burdens them w ith serious drawbacks fo r systematic
studies.
Because the prim ary focus o f psychotherapy is on people and th e ir
problem s, n o t on dreams, the actual report o f the dream may be transform ed
in the process o f discussing it or n o t w ritte n dow n in fu ll d e ta il by the
therapist u n til after the session or la ter in the day. M oreover, the dreams
reported by clients tend to focus on the issues in the fo re fro n t o f the
therapeutic relatio nship, as shown in a study com paring dreams reported
durin g awakenings in a sleep laboratory and those to ld to the particip ants’
psychotherapist (W h itm a n , Kram er, & Baldridge, 1963). A s a result, the fu ll
range o f the clie n ts’ dream life is n o t covered, leading to an unrepresentative
sample. T hen, too, the dreams analyzed in published case studies are n o t
usually presented in fu ll detail, m eaning th a t la te r use o f them is im possible
w ith o u t co n ta ctin g the therapists, who are ofte n concerned about p rotecting
the privacy o f th e ir clients. M ost damaging o f a ll, th e demand characteristics
o f a psychotherapy relationship are far greater th a n those o f an experim ental
setting, a p o in t th a t is elaborated la te r in this chapter in a discussion o f
free association as a m ethod fo r analyzing dream content.
In summary, w hen the advantages and disadvantages o f the various
m ethods o f o b ta in in g dream reports are weighed, it seems th a t m ost recent
dreams and dream journals provide the best cu rre n tly available sources. T he
m ost recent dream m ethod is m ore structured, standardized, and faster than
the dream diary m ethod, m aking it possible to obtain dreams from more
members o f a group th a n does the dream diary approach. T he use o f in d iv id u a l
dream journals enables the study o f a larger num ber o f dreams than otherwise
w ould be feasible. I t also provides a m ore representative sample o f a person’s
dream life th a n the dreams obtained from a person in psychotherapy.
T H E REPR ESENTATIVEN ESS O F D R E A M R EC ALLER S
There are m any people w ho do n o t recall even one dream a week and
therefore are less lik e ly to co n trib u te dream reports, regardless o f the m ethod
o f co lle c tio n . Is it th e n possible to generalize about dream co n te n t o n the
basis o f those people w ho are able and w illin g to provide dream reports?
For personality variables, the answer has been in a d ve rte n tly provided by a
large num ber o f studies stretching over several decades. T he studies a t
tem pted to fin d personality differences between those w ho recall dreams
and those w ho do n o t, as determ ined by responses to a sim ple questionnaire
o r the to ta l num ber o f dreams entered in to a 2-week dream diary.
T he m ost im p o rta n t fin d in g from the studies, w h ich used a w ide range
o f personality tests, is th a t no consistent differences exist between recallers
and nonrecallers (D om hoff, 1996, chapter 3 ): O n the personality dim ensions
th a t psychologists can measure, study participants w ho co n trib u te d dreams
were sim ila r to those w ho d id n o t (B errien, 1933; C ohen, 1979; D o m h o ff
& G erson, 1967; Farley, S chm uller, & Fischbach, 1971; S ticke l, 1956;
Tonay, 1993; T rin d e r & Kram er, 1971). C ohen (1979), w ho conducted
several exce lle n t studies on th is question, provided a good summary o f the
lite ra tu re w hen he concluded: “ C orrelations between dream re ca ll frequency
and specific personality measures have been weak, triv ia l, o r in consisten t”
(p. 161). Based on th e ir ow n findings 21 years after C ohen w rote, along
w ith a review o f the m ore recent lite ra tu re , Blagrove and A ke h u rst (2000)
reached the same general conclusion.
O n the o th e r hand, co g n itive variables seem to have a s lig h tly greater
co rre la tio n w ith dream recall th a n do personality variables (C ohen, 1979;
C ory, O rm isto n , S im m el, & D a ino ff, 1975; F itc h & A rm itage, 1989;
H a rtm a nn et al., 1991; H iscock & C ohen, 1973; M a rtin e lli, 1983). T h is is
especially th e case fo r visuospatial a b ility , w h ic h was investigated in adults
w ho had little or no dream recall by B u tler and W atson (1985) as a fo llo w
up to the developm ental findings by Foulkes (1982) discussed in chapter 1.
P hysiological factors unrelated to personality or co g n itive variables
also seem to play a p a rt in m aking some people less able to recall th e ir
dreams, as seen in a study show ing th a t some low recallers have h ig h w aking
thresholds and are d iffic u lt to arouse in a sleep laboratory (Zim m erm an,
1970). R ecall frequencies also vary w ith m ood and stress levels (A rm itag e,
1992; C ohen, 1979). S till, none o f the various factors m entioned up to th is
p o in t seems as im p o rta n t as an interest in dreams (C ohen & W o lfe , 1973;

Strauch, 1969), a fin d in g th a t does n o t correlate w ith personality variables
(T onay, 1993) and is supported by the fact th a t less v a ria tio n in recall
usually occurs among adu lt participants w ho are studied in sleep laboratories.
G enerally speaking, then, there seems to be no one reason w hy some
people do n o t recall dreams w ith any degree o f regularity (S chredl & M ontas-
ser, 1996, 1997). G oodenough’s (1991) assessment o f this lite ra tu re summa
rizes the issue w e ll: “ People apparently may be non-reporters fo r a va rie ty o f
d is tin c tly d iffe re n t reasons” (p. 157). I t therefore follow s th a t a representative
sample o f people is able to co n trib u te to studies o f dream content. T h a t is,
fo r purposes o f group comparisons, dream recallers provide diverse and
representative samples on personality and cogn itive factors if an adequate
sample is available. A s dem onstrated in chapter 3, “ adequate sample size”
means at least 125 dream reports per sample because m any o f the elements
in dreams appear in h a lf or less o f a ll dream reports.
These findings have another im p lic a tio n as w ell: T he y suggest th a t
fo r purposes o f studying the relatio nship between dream con te n t and w aking
c o g n itio n outside the laboratory, it m ig h t make sense to focus on people
w ho are good dream recallers. T h is atypical strategy gains p la u s ib ility w ith in
the co n te xt o f the m any general findings on dream con te n t th a t are presented
in a previous study (D om hoff, 1996) and throughout this book. A research
m ethodology fo r using unique participants w ith good recall skills is presented
and dem onstrated in chapter 5.
M E T H O D S FO R A N A L Y Z IN G D R E A M C O N T E N T
O nce dream reports have been collected, four m ethods may be used
to analyze dream content. T he methods range from the subjective to the
objective, and vary in the num ber o f dreams w ith w h ich they may be used.
A s is the case w ith co lle ctin g dreams, each m ethod o f analysis has strengths
and weaknesses. In the present stage o f dream research, however, the rela
tiv e ly objective and q u a n tita tive m ethods seem to have the greatest p o te n tia l
fo r testing hypotheses o rig in a lly developed on the basis o f in d iv id u a lis tic
and subjective m ethods. In the order o f th e ir appearance in the dream
lite ra tu re , the m ethods are (a) free association, w h ich led Freud (1900) to
the conclusions on w h ich he b u ilt his w is h -fu lfillm e n t theory o f dreams;
(b ) sym bolic in te rp re ta tio n , w h ich was one basis fo r Jung’s (1963) break
w ith Freud and the developm ent o f his ow n theory; (c) them atic analysis
o f dream journals, w h ich involves a search fo r repeated topics, a ctivitie s,
o r events; and (d) con te n t analysis, w h ich involves the con stru ctio n o f
ra tin g scales or n o m in a l categories to study large numbers o f dreams from
e ith e r groups or individuals.
Free Association
T he free-association m ethod consists o f in stru ctin g dreamers to say

w hatever comes in to th e ir m inds about each elem ent o f the dream w ith o u t
any censoring o f th e ir thoughts. A chain o f free associations can be long
and can lead to seem ingly unrelated topics. O nce the free associations are
obtained fo r a ll parts o f the dream, they are organized to in fe r the underlying
m otives fo r the dream its e lf (th e latent content, in Freud’s term s). A lth o u g h
used p rim a rily by Freudians, the m ethod can be used by non-Freudians
because it ofte n reveals the day-to-day events incorporated in to the dream
and the em otion al concerns o f the dream er (C ip o lli & P o li, 1992).
T he free association m ethod fo r system atic research has m ajor prob
lems. M ost generally, it is im possible to be sure w ith in a psychotherapy
setting w hether the free associations actually explain the dream, because
so m uch else is know n about the dream er th a t could be playing a role in
constructing a “ m eaning” fo r the dream. Even more seriously— and despite
claim s by psychotherapists th a t they m a in ta in a neu tral and nonjudgm ental
stance— it is d iffic u lt to rule out an a lte rn a tive analysis o f dream in te rp re ta
tio n developed by social psychologists and m emory researchers. T he basic
idea is th a t therapists u n kn o w in g ly shape th e ir clie n ts’ associations through
com plex processes o f suggestion, persuasion, and conversion.
T he social psychology o f the psychotherapy relatio nship begins w ith
the c lie n t’s desire fo r help and a great respect fo r the expertise o f the
therapist. In th a t co n te xt, a dream in te rp re ta tio n by the therapist th a t comes
as a surprise or seems u n lik e ly to the c lie n t leads h im or her in to a state
o f u n ce rta in ty and co g n itive dissonance. G radually, how ever, the c lie n t
comes to agree w ith the therapist about the m eaning o f the dream through
th e tendency to obey a u th o rity figures and a desire to reduce co g n itive
dissonance (Loftus & Ketcham , 1994; O fshe & W atters, 1994). T h is se
quence o f events is reinforced if details in la te r dream reports provide an
occasion fo r resolving the co g n itive dissonance. Because dreams are generally
th o u g h t to be beyond the reach o f suggestion, such confirm ato ry dreams
can play a pow erful role in a conversion process.
A s a result o f th is sequence o f events, the person takes o n a new
id e n tity as “ a c lie n t o f a famous therapist” o r as a member o f “ an exclusive
school o f th o u g h t.” In effect, the person is “ cured” by ta kin g on a new
id e n tity th a t includes an in-group language fo r understanding the w orld,
attendance at events w ith fe llo w converts, and even becom ing a p ra c titio n e r
o f the therapy. In o ther words, the explana tion developed by social and
co g n itive psychologists fo r w hy dream analysis and hypnosis can lead to
“ recovered m em ories” and “ b e lie ve d -in im aginings” is applicable to the
in te rp re ta tio n o f dreams w ith in the psychotherapy process (A y e lla , 1998;
de R ivera & S arbin, 1998). M oreover, experim ental evidence fro m sim ulated

dream interpretations supports th is in te rp re ta tio n o f w hat can happen in
psychotherapy. In those studies, a c lin ic a l psychologist convinced many
participants th a t they had the experience o f being lost before age 3 (M azzoni
& Loftus, 1998; Mazzoni, Loftus, Seitz, & Lynn, 1999). U n til c lin ic a l theo-
rists can answer the foregoing analysis and evidence by showing th a t sugges
tio n is n o t operating in dream in te rp re ta tio n , the usefulness o f free associa
tio n as an objective m ethod fo r understanding dreams in psychotherapy
remains in question.
T he usefulness o f free association outside o f psychotherapy also is in
doubt. Foulkes’s (1978) extensive e ffo rt to use the m ethod as a research
to o l w ith dreams collected in the laboratory d id n o t prove to be successful
in creating a “ gram mar o f dreams.” Instead, Foulkes (1996a) la te r w rote
th a t “ extensive experience in association gathering” convinced h im o f the
“ in h e re n t arbitrariness” o f the m ethod (p, 617). For example, a dream er’s
free associations w ould fa il to m e n tio n d ire ct, re a l-life parallels to the dream
th a t were know n independently to the person co lle ctin g the free associations.
M oreover, tw o studies published 35 years apart were unsuccessful in establish
ing the increm ental v a lid ity o f free associations in b lin d analyses o f dream
series containin g 15 to 25 dreams. B o th studies found th a t the analysts who
had free associations along w ith the dream series made no m ore correct
inferences th a n did those who only had the dream series (Popp, Luborsky,
& C rits-C h risto p h , 1992; Reis, 1959). Thus, it seems u n lik e ly th a t free
associations can be considered a useful to o l fo r systematic analyses o f dream
co n te n t outside the c lin ic a l situatio n. T he m ethod proved arbitrary in one
study and superfluous in tw o others.
M etaphoric A nalysis
“Sym bolic” interpretations are used as a supplem ent to free associations

in psychotherapy settings and in some studies o f long-term dream journals.
T hey are the essence o f Jung’s (1974) “ a m p lifica tio n ” m ethod (see also
M attoon , 1978). T h is m ethod differs from free association in th a t b o th the
dreamer and the dream in te rp re te r produce a w ide range o f images and
associations d ire c tly related to the dream. T he y do so by re tu rn in g to the
dream after each new though t o r image, ra th e r th a n having one association
lead to the next. In lig h t o f the w o rk o f H a ll (1953a), L a ko ff (1993a, 1997),
and States (1987), it seems clear th a t these sym bolic interpretations are
now more appropriately though t o f as m etaphoric analyses.
In fact, it can be argued th a t the riv a l c lin ic a l schools o f sym bolic
in te rp re ta tio n are particular applications o f com m on m etaphoric under
standings to aspects o f dreams, as the w o rk o f the neo-Freudian E rich From m
(1949, 1951) demonstrates. Even the analyses made by phenom enological
and e xiste n tia l theorists, such as Boss (1958, 1977) and Peris (D ow ning &
l
M arm orstein, 1973), w ho cla im to re je ct sym bolic in terpretations, are actu
a lly m etaphoric glosses o f dreams, as can be seen by reading th ro u g h th e ir
case examples.
T here are several problem s w ith m etaphoric analyses, sta rtin g w ith a
fa ct m entioned in chapter 1: N o system atic evidence indicates w hether
dreams are m etaphoric in nature. Even i f some aspects o f dream co n te n t
are m etaphoric, no guidelines exist as to w h ich o f m any possible conceptual
m etaphors should be applied. A lth o u g h H a ll (1953a) argued th a t th e re p e ti
tio n o f elem ents in a dream series can lead to plausible evidence fo r applying
one o r another conceptual m etaphor, and he provided some rough guidelines
fo r id e n tify in g possible m etaphors in dreams, m etaphoric analysis as a rig o r
ous and system atic approach rem ains undeveloped.
Them atic Analysis
T he th ird m ethod o f dream analysis, the them atic m ethod, shades o ff

from m etaphoric analysis. I t involves reading through a dream series several
tim es to see i f ce rta in settings, objects, or events appear several tim es. T he
various possibilities are seen as pieces o f a puzzle th a t m ust fit together
before any te n ta tive analysis is taken m ore seriously. Sometimes the search
fo r themes is made easier by the presence o f one or m ore seem ingly obvious
o r “ barefaced” dreams, w h ich H a ll (1947) called spotlight dreams. S p o tlig h t
dreams o fte n co n ta in several key topics in the series in a d ire ct fashion.
Inferences th a t are based on sp o tlig h t dreams are th e n exam ined by going
back thro u g h th e dream series to see if less obvious dreams tu rn o u t to have
a sim ila r structure.
A study using th is approach o n 649 dreams over a 50-year period found
th a t six themes appeared w ith the same frequency thro ugho ut the dream
series (H a ll & N ordby, 1972; S m ith & H a ll, 1964). T he dream er was
th in k in g about, preparing, o r eating food in 20% o f her dreams; m isplacing
an object, usually her purse, in 17%; fin d in g herself in a sm all o r disorderly
room in 10%; in te ra ctin g w ith her m other in 10%; going to the bathroom
in 8%; and being late o r m issing a bus or tra in in 6%. As m entioned in
chapter 1, these six themes account fo r at least p a rt o f the co n te n t in 71%
o f the dreams.
A lth o u g h it is a little easier fo r investigators to reach agreem ent on
the presence o f themes th a n it is on m etaphors, considerable room fo r
disagreem ent remains. T he m ethod also suffers because the findings tend
to be unique to each dreamer, a llow ing little o p p o rtu n ity fo r generalizations
across dreamers o r groups o f dreamers. F in a lly, them atic analyses tend to
be general. T hey do n o t go far in p ro vid in g detailed statem ents about dream
co n te n t th a t can be tested on new dream samples or correlated w ith the
effects o f b ra in lesions or drugs.

C ontent A nalysis
D issatisfaction w ith the re lia b ility and generalizability o f free associa

tiv e , sym bolic, and them atic m ethods o f studying dream co n te n t led to the
use o f the m ore objective and q u a n tita tive approach called content analysis.
C o n te n t analysis is a general m ethod th a t attem pts to use carefully defined
categories to extract m eaning from a te xt, such as a newspaper article,
transcribed conversation, short story, o r dream report. O ne o f the earliest
proponents o f con te n t analysis stated th a t the “ fundam ental objective” o f
th is m ethod is to convert the “ sym bolic behavior” o f people in to “ scie ntific
data,” by w h ich he m eant data th a t were (a) objective and reproducible;
(b ) susceptible to measurement and q u a n tifica tio n ; (c) sig n ifica n t fo r either
pure or applied theory; and (d) generalizable (C a rtw rig h t, 1953, p. 466).
H a ll (1969a) defined con te n t analysis as “ the categorization o f units o f
q u a lita tive m aterial in order to obtain frequencies w h ich can be subjected
to sta tistica l operations and tests o f significance” (p. 175). Thanks to the
developm ent o f sophisticated software, content analysis now includes pro
grams fo r aggregating words and phrases in to “ sem antic netw orks” th a t do
n o t rely on predeterm ined categories (Roberts, 1997).
T he m ost d iffic u lt task in carrying o u t a con te n t analysis is to develop
categories th a t lead to reliable and va lid findings. U n fo rtu n a te ly, there are
no general rules fo r constructing such categories, n o r has it been found th a t
categories created fo r one type o f te x t can be readily used w ith texts o f
another k in d . For the m ost part, co n te n t categories have been developed
through tria l and error after fu ll im m ersion in the type o f te x t to be analyzed.
T hey usually go through several versions before they are ready fo r regular use.
T w o m ain issues are in vo lve d in developing co n te n t categories: deter
m in in g the le ve l o f measurement and deciding w hether to use em pirical or
th eoretical categories. C oncerning the le vel o f measurement, the choice is
basically between ra tin g scales at the o rd in a l level, w h ich im p ly the existence
o f degrees o f difference th a t can be ranked, and discrete categories at the
no m in a l level. “ A c tiv ity le ve l” and “ em otional in te n s ity ” are examples o f
dim ensions th a t have been used on ra tin g scales fo r the study o f dreams.
Indoor and outdoor settings and male and female characters are examples
o f categories at a nom in al le ve l o f measurement, where a sim ple ta b u la tio n
o f frequencies is made fo r each category.
Em pirical categories are based on w hat appear to be natural dim ensions,
or clusterings th a t derive from hum an experience, w ith o u t regard to any
p a rticu la r psychological theory. As noted in chapter 1, the categories are
usually basic-level cogn itive categories. “ Characters,” “ social in teractio ns,”
and “ vividness” are examples o f em pirical categories th a t seem to fit w ith
everyday understandings. Theoretical categories, on th e other hand, bring
together seem ingly disparate and unconnected elements in dream reports
on the basis o f a careful rendering o f a concept fro m a theory o f in terest
to the investigator. “A n im a ,” “ castration an xie ty,” and “ ego synthesis” are
examples o f th e o re tica l scales developed fo r the study o f dream reports
from Jungian, Freudian, and E riksonian theory, respectively (H a ll, 1969a;
Sheppard, 1969).
T he ra tin g -n o m in a l and e m p iric a l-th e o re tic a l dichotom ies lead to
the p o ssib ility o f fo u r d iffe re n t types o f scales, and in fact, a ll fo u r types have
been used in dream research, some m ore freq uently th a n others. G enerally
speaking, m ost o f the coding systems fo r the study o f dreams have been
e m p irica l ra tin g scales. A fa cto r analysis o f the codings o f 100 R E M dream
reports w ith several d iffe re n t em pirical scales suggests th a t these scales b o il
dow n to five basic dim ensions: degree o f vividness and d is to rtio n ; degree
o f a nxiety and h o s tility ; degree o f in itia tiv e and strivin g ; le ve l o f a c tiv ity ;
and am ount o f sexuality (H a u ri, 1975). I t also can be said th a t some types
o f scales have been m ore useful. In particular, e m p irica l scales, w hether at
the n o m in a l o r o rd in a l level, have proven to be m ore useful th a n e ith e r
type o f th e o re tica l scale. T h e o re tica l scales are d iffic u lt to construct as a
result o f the fuzziness o f m ost personality theories, and they are equally
d iffic u lt to apply because o f the am biguity o f m any dream actions (H a ll,
1969a). D o m h o ff (1996, p. 10) provides details o n the fa ilu re o f tw o scales
developed to test Freudian ideas.
Rating Scales fo r Dream Content
R a ting scales, as already noted, are based on th e assum ption th a t a

characteristic can be ranked or weighed. A ll ratings scales in dream research
have been at the o rd in a l le ve l and have rested on th e assum ption th a t “ m ore”
or “ less” is the m ost th a t can be judged in a dream report. Nevertheless, one
group o f theorists assigned w idely d iffe re n t w eights to the points along
th e ir psychoanalytic scales fo r ego fu n ctio n in g , a decision hard to ju s tify in
measurement term s (Sheppard, 1963, 1969). O rd in a l scales have been used
w ith great benefit in a w ide va rie ty o f studies, the m ost im p o rta n t o f w h ic h
are the lo n g itu d in a l and cross-sectional studies by Foulkes and his co-workers
(1982; Foulkes et al., 1990). T he scales used in th e ir w o rk made it possible
to dem onstrate the system atic changes in dream co n te n t from the preschool
to teen years th a t are discussed in chapter 1.
R a ting scales are m ost useful fo r characteristics o f dream reports w ith
degrees o f in te n s ity in w aking life , such as a c tiv ity le vel or e m o tio n a lity,
or w ith o u t specific content, such as c la rity o f visual im agery or vividness.
Som etim es useful ratings on these kinds o f dim ensions are made by the
dreamers themselves. For exam ple, Foulkes (1966) used ratings by b o th
judges and particip ants on dram atic q ua lity, degree o f unpleasantness, and
c la rity in show ing th a t few differences exist in the dream reports from the

first three REM periods o f the n ig h t. In a sim ilar fashion, Howe and B lic k
(1983) had wom en rate th e ir dream reports on several e m o tio n a lity dim en
sions, fin d in g th a t the older wom en gave m ore benign ratings to th e ir
dream em otions.
A cco rd in g to C ohen (1979), four dim ensions o f dream salience can
be rated by participants in dream studies: e m o tio n a lity, bizarreness, a c tiv ity ,
and vividness. These dim ensions are somewhat sim ila r to the findings from
H a u ri’s (1975) facto r analysis o f rating scales, w h ich also p in p o in te d v iv id
ness and a c tiv ity dim ensions; H a u ri’s factor o f a n x ie ty -h o s tility may be
sim ila r to C ohen’s e m o tio n a lity dim ension as w e ll. C ohen’s conclusion is
supported in a study using m any pairs o f polar adjectives and 7 -p o in t L ik e rt
scales to establish correlations between levels o f EEG a c tiv ity and im m ediate
ratings o f dream experiences by laboratory participants (T akeuchi, O g ilvie ,
F e rre lli, M urphy, & B e licki, 2001). A facto r analysis showed th a t the Dream
P roperty Scale (T akeuchi et al., 2001) derived from these polar adjectives
has the fo llo w in g fo u r m ain factors:
1. E m otion ality/evaluatio n, as indexed by pairings such as re

laxed-tense and pleasant-unpleasant.
2. R ationality/bizarreness, as indexed by pairings such as fa m il
ia r-u n fa m ilia r and ordinary-strange.
3. A c tiv ity , as indexed by pairings such as alert-dro w sy and
dyna m ic-static.
4. Im pression/vividness, as indexed by pairings such as clear-fuzzy
and focused-unfocused.
I t is notew orthy th a t tw o o f the dim ensions, E m o tio n a lity/e va lu a tio n and

A c tiv ity , are sim ila r to tw o o f the three dim ensions th a t have been found
to be universal to v irtu a lly a ll kinds o f ratings o f affective m eaning in m any
d iffe re n t language fam ilies across the w o rld (Osgood, M ay, & M iro n , 1975).
T h e th ird universal dim ension o f affective m eaning, potency, w h ic h is best
indexed by the polar adjectives strong-w eak, may have some overlaps w ith
the R ationality/bizarreness dim ension o f the Dream Property Scale.
Despite these examples o f useful applications, there are nonetheless
serious drawbacks to ra tin g scales w hen it comes to detailed studies o f dream
content. First, it is o fte n d iffic u lt to establish re lia b ility w ith some scales,
especially w hen researchers from outside the o rig in a l investigative team try
to use them (D om hoff, 1996; W in g e t & Kram er, 1979). P a rtly fo r th is
reason, new investigators tend to create th e ir ow n ra tin g scales, leading to
a situ a tio n in w h ich m any scales have n o t been fu lly tested fo r e ith e r
re lia b ility or v a lid ity . For example, in a study o f the possible differences
between the dreams o f people who d iffe r on the perm eability o f th e ir personal
boundaries, three o f the new scales used in the study had to be excluded
fro m the analysis because o f low intercoder re lia b ilitie s (H artm ann , Rosen,
&. Rand, 1998). T h is constant creation o f new scales also means th a t results
cannot be d ire c tly compared from study to study, m aking it d iffic u lt to b u ild
a solid and reliable research lite ra tu re .
Second, m uch o f the specific in fo rm a tio n in dream reports can be lost
or unused w ith ra tin g scales. A “ bizzareness” scale, fo r instance, does n o t
include the fa ct th a t in one set o f dream reports the h ig h degree o f bizarreness
may be due to metamorphoses, in another to im possible actions by specific
dream characters, and in s till another to im plausible settings or objects.
S im ila rly, the highest ra tin g on a h o s tility scale may be the result o f e ith e r
a m urder o r a fa ta l illness, b u t the difference between the tw o may be as
in fo rm a tive as the e xtre m ity o f the situatio n.
T h ird , m any ra tin g scales rest on assumptions th a t are psychologically
untenable w hen they are exam ined c ritic a lly . For exam ple, in a dependency
ra tin g scale created by W h itm a n , Pierce, Maas, and Baldridge (1961), a
score o f 6 is assigned i f the person eats food and a score o f 1 is assigned if
the person seeks he lp from others. Because the ratings fo r each dream are
added together to create the to ta l score, th is ra tin g system im p lic itly assumes
th a t “ m e n tio n in g a ham sandw ich shows six tim es as m uch dependency as
accepting a h e lp in g hand from another” (V a n de Castle, 1969, p. 193).
T h is type o f psychologically untenable assum ption is m ost prevalent
in ra tin g scales fo r aggressive actions. W ith m ost o f these scales, murders
receive the highest ratin g ; in ju ry and damage to personal possessions receive
m edium scores; and insults, rejections, and expressions o f h o s tility receive
low scores. T he ratings fo r each dream are added together, and an average
aggression score is calculated fo r each in d iv id u a l or group. Such a procedure
im plies th a t several angry thoughts or a few damaged possessions are psycho-
lo g ica lly equ ivalent to one m urder, a w e ig h tin g th a t seems indefensible once
it is made e x p lic it (H a ll, 1969a, 1969b). A ra tin g scale fo r levels o f anger
is reasonable, b u t a ra tin g scale fo r aggressions th a t range from insults to
m urders is n o t. T h is problem is one o f the m ajor failings w ith the G ottschalk
and Gleser (1969) scales, w h ich were o rig in a lly developed to study psycho
therapy records and o th e r form s o f w aking ta lk. Despite th e ir use in several
dream studies, the scales do n o t adapt w e ll to the study o f dream content.
D o m h o ff (1996, pp. 3 0 -3 7 ) presents critiques o f several o th e r inadequate
ra tin g scales.
F ourth, and fin a lly , ra tin g scales are usually n o t o f m uch use in tryin g
to relate the substance o f dream co n te n t to w aking conceptions, concerns,
and interests. T h e y are best used fo r re la tin g features o f dream ing to the
neurophysiology o f sleep. Thus, as useful as the Dream Property Scale is
fo r re la tin g perceived E m o tio n a lity/e va lu a tio n o r A c tiv ity le ve l to EEG
variables, it cannot be used to study the m eaning th a t m ig h t be found in
dream reports.

Nom inal Scales
N o m in a l scales do n o t suffer from the problem s facing ra tin g scales.

H ig h e r re lia b ilitie s can be obtained because discrete scales usually are more
clearly defined and seldom require the subtle judgm ents th a t ra tin g scales
often do. N o in fo rm a tio n is lost because numerous categories can be created
and th e n aggregated, if necessary, la te r in the data analysis. N o m in a l catego
ries do n o t con ta in the questionable psychological assumptions b u ilt in to
some ra tin g scales. Nonetheless, n o m in a l-le ve l coding systems are n o t w ith
out th e ir d ifficu ltie s and drawbacks, w h ich result in less use o f them than
m ig h t be expected.
First, it takes tim e to create carefully defined categories th a t can lead
to h ig h intercoder re lia b ility , and there may be some gray areas even w ith
w ell-defined categories. Second, the categories may n o t prove to be useful
because they are n o t as lik e ly to be created on the basis o f in tu itio n s about
the data. M ost im portant, a good coding system at the no m in a l le vel is far
more labor intensive th a n ra tin g systems. I t takes m any hours to learn a
fu ll set o f n o m in a l categories and th e n m any m ore hours to apply them to
a sample o f dreams th a n is the case w ith m ost ra tin g systems.
The H a ll-V a n de Castle coding system used in th is book is atypical
in the area o f dream research in th a t its categories are no m in a l in nature.
T he system o rig in a lly contained b o th em pirical and th e o re tica l categories,
b u t the th e o re tica l categories did n o t prove to have any more usefulness or
v a lid ity than did the th eoretical ra tin g scales, and they have long since
been abandoned (D om hoff, 1996). T he o rig in a l H a ll-V a n de Castle system
consists o f eight general categories, m ost o f w h ich are divided in to tw o or
m ore subcategories. T he categories encompass the five dim ensions found in
ra tin g scales by H a u ri’s (1975) fa cto r analysis. As noted earlier, this system
is discussed in d e ta il in chapter 3.
D E T E R M IN IN G A U N IT O F A N A LY S IS
W h ether ra tin g scales or n o m in a l categories are used, it is necessary

to decide on the u n it o f analysis to be used in m aking standardized com pari
sons from dream sample to dream sample. First, and m ost crucially, wide
in d iv id u a l differences exist in re p o rt length, and w om en’s dreams o fte n are
found to be longer th a n those o f m en (B ursik, 1998; H a ll &. V a n de Castle,
1966; W inegar & Levin, 1997). V arying lengths are a problem because
longer reports are lik e ly to have m ore o f m ost things in them . Second,
dream reports can vary from group to group o r person to person in the
frequency w ith w h ich certain elem ents appear, even w hen report length is
held constant. T h is difference in “ density” seems to be especially the case
fo r the frequency o f characters, m eaning th a t some dream reports are m ore
lik e ly to co n ta in social interactions th a n others are. Once again, there is a
gender difference: M ore characters exist in w om en’s reports, an in terestin g
fin d in g in and o f itse lf, but one th a t should be taken in to account in analyzing
social interactions (H a ll, 1969a, 1969b).
T he fa ilu re to correct fo r dream le n g th is a problem w ith b o th ra tin g
scales and n o m in a l categories. For instance, a freq uently used th e o re tica l
ra tin g scale fo r “ prim ary process th in k in g ” in dream content, w h ic h requires
d iffic u lt judgm ents concerning d iffe rin g degrees o f d is to rtio n and im p ro b a b il
ity , correlates .60 w ith the le n g th o f the dream report (A u ld , G oldenberg,
&. W eiss, 1968). W h e n controls fo r dream le n g th are included, the previously
reported positive relationships between th is scale and c re a tiv ity measures
disappear (L ivin g sto n & L e vin , 1991; W ood &. D om ino, 1989). S im ila rly,
co rre ctio n fo r dream le n g th elim inates seeming gender differences in several
H a ll-V a n de Castle categories.
T he fa ilu re to c o n tro l fo r le ngth is one o f several m ethodological
problem s in a study by Kram er, K inne y, and S charf (1983), w h ich w rongly
claim ed th a t previous gender differences reversed or disappeared; in th e ir
sample, the m en’s dream reports were longer th a n those o f the wom en (see
D om hoff, 1996, pp. 7 9 -8 2 , fo r a fu ll c ritiq u e ). In a study com paring dream
reports fro m wom en w ho w orked in the hom e w ith those from wom en and
m en who worked outside the hom e, the few differences cannot be regarded
as solid findings because no corre ctio n was made fo r the fact th a t the dreams
o f the wom en w ho w orked at hom e averaged 220 words, those o f the w om en
w ho worked outside the hom e 200 words, and those o f the m en o n ly 180
words (Lortie-Lussier, Sim ond, R in fre t, & De K o n in c k , 1992).
T w o m ain strategies have been used to correct fo r d iffe rin g lengths o f
reports. First, the mean num ber o f lines o r words per dream report was used
as the u n it o f analysis. H ow ever, th a t strategy does n o t deal w ith the d iffe rin g
“ wordiness” o f particip ants and leads to cumbersome findings such as “ there
were 2.3 hum an characters per every 10 lines (o r 100 words) in the dream
narratives.” W orse, a c o n tro l fo r le ngth th a t relies on the num ber o f words
does n o t allow fo r com plicated o r unusual elem ents th a t m ig h t take more
words to describe; consequently, d iv id in g by the num ber o f words could
wash o u t real and im p o rta n t differences, especially in studies concerned
w ith c re a tiv ity o r unusual features in dreams (H u n t, R uzycki-H unt, Pariak,
& B e licki, 1993).
Second, investigators established m inim u m and m axim um lengths fo r
the reports to be analyzed, thereby m aking it possible to use the dream
report as a w hole as the u n it o f analysis. T h is approach was used by H a ll
and V a n de Castle (1966) w hen they elim in a te d reports o f fewer th a n 50
o r more th a n 300 words in a norm ative study o f dreams from college m en
and wom en (see chapter 3). T h is is, in fact, the o n ly approach possible

w ith ra tin g scales because ratings are based on the dream as a w hole. In
effect, th is lim ita tio n is another disadvantage o f using ra tin g scales to study
dream content.
There is, how ever, a far b e tte r way to create sensible u n its o f analysis
w ith n o m in a l categories, w h ich is to convert categorical frequencies in to
percentages and ratios. T h is approach is now used w ith the H a ll-V a n de
Castle system, as shown in chapter 3. For example, it is possible to make
dreams from groups a ll over th e w o rld com parable w ith regard to th e ir
in clu sio n o f animals by creating an “ anim al percent.” T he anim al percent
is sim ply the to ta l num ber o f anim als in the dreams divide d by the to ta l
num ber o f a ll types o f characters. A s shown w ith systematic data in chapter
3, th is approach is independent o f report le ngth or character density w ith in
broad lim its ; it e ffe ctively deals w ith b o th problem s at the same tim e. In
a d d itio n , findings presented as percentages and ratios are readily com m uni
cated and com prehended. People im m ediately understand if it is reported
th a t the anim al percent in dreams declines from the 30% to 40% range in
ch ild h o o d to 4% to 6% in adulthood and is h ig h e r in the sm all tra d itio n a l
societies studied by cu ltu ra l anthropologists th a n it is in industrialized democ
racies (D om hoff, 1996; V a n d e C astle, 1983).
Percentage indicators also make it possible to resolve an im p o rta n t
controversy over the way to compare REM and N R E M reports. W hen
lengths are held constant, the seeming differences between the tw o types
o f reports disappear (A ntrobus, 1983; Foulkes & S chm idt, 1983). Plowever,
c ritic s argue th a t no correction should be made because N R E M m entation
is, in fact, o f briefer duration and more th o u g h tlike th a n REM m enta tio n
(H obson et al., 2000b, p. 800; H u n t, 2000). U sing 125 R E M and 55 N R E M
reports collected by Foulkes and Rechtschaffen (1964), a reanalysis th a t
was based on percentage indicators fo r a ctivitie s in dreams was used as a
test case. A lth o u g h 87% o f the R E M reports and 89% o f the N R E M reports
had at least one a c tiv ity in them , the p ro p o rtio n o f a ll a ctivitie s th a t were
co g n itive activitie s was 20% in N R E M reports, but o n ly 11% in REM
reports. Conversely, higher visual and verbal proportions occurred in REM
th a n in N R E M reports: 12% versus 6% fo r visual, 37% versus 22% fo r verbal
(D o m h o ff &. Schneider, 1999. p. 149). Thus, differences between REM
and N R E M reports persist w hen percentage indicators are used to co n tro l
fo r length.
T H E PROBLEM S OF S T A T IS T IC A L A N A LY S IS
T h e nature o f the data in m ost studies o f dream co n te n t leads to

d iffic u lt statistical problems th a t c a ll in to question published claim s th a t
are based on param etric tests lik e the t test and analysis o f variance. First,
statistics textbooks generally argue th a t it is risky to use param etric statistics
w ith o rd in a l o r no m in a l data because such statistics require th a t the points
along any m easurement scale reflect an underlying continuous d is trib u tio n
w ith at least equal in tervals (e.g., Leach, 1979; Reynolds, 1984; Siegel &
C astellan, 1988). I t is n o t im possible to add, subtract, m u ltip ly , o r d ivid e
the frequencies derived from n o m in a l categories; it is ju st th a t such an
analysis may be m isleading or in error because assumptions are being vio lated.
A s Siegel and C astellan (1988) stressed:
It should be obvious that a mean and standard deviation may be com

puted for any set of numbers. However, statistics computed from these
numbers only “ make sense” if the original assignment procedure im
parted “arithm etical” interpretations to the assignments. This is a subtle
and critica l point, (p. 33)
A lth o u g h evidence has indicated th a t using param etric statistics usually

is n o t fa ta l i f d istrib u tio n s are close to norm al (Saw ilow sky (St B la ir, 1992),
it is ce rta in th a t m ost o rd in a l and no m in a l data collected on people outside
a laboratory setting are n o t n orm ally distributed, a s itu a tio n th a t violates
one o f the m ost im p o rta n t underlying assumptions fo r the a p p lica tio n o f
param etric statistics (M ic c e ri, 1989). U n der these circum stances, w h ich
apply to m ost o f the elements th a t appear in dream reports, nonparam etric
statistics have been shown to w ork m uch better, especially w hen sample
sizes are uneven. E m pirical studies show th is approach to be true even w ith
the larger sample sizes th a t are though t to favor param etric statistics (N anna
&. Sawilow sky, 1998). T h ird , the standard form ulas fo r d eterm ining p values
w ith param etric statistics assume th a t the findings are based on random
samples, w h ich are rarely obtained in dream research.
T h e best way to deal w ith these problem s is to use the co m b in a tio n
o f nonparam etric statistics and random ization strategies explained in chapter
3. T h is approach is especially w e ll suited fo r the n o m in a l data produced by
the H a ll-V a n de Castle coding system. I t also elim inates any need to w orry
about the le vel o f measurement, the randomness o f samples, or the shape
o f the sam pling d is trib u tio n . In a dd ition , it elim inates the need fo r equal
sample sizes and makes it unnecessary to use d iffe re n t statistics fo r analyzing
lo n g itu d in a l data from in d iv id u a l dream journals.
T h e Issue o f Sample Sizes
S ta tis tic a l problem s in dream research are. com pounded by th e frequent

use o f sample sizes th a t are far too sm all to allow fo r the d e te ctio n o f
sta tistica lly sig n ifica n t differences. There is a tendency fo r researchers to
use o n ly from 30 to 60 dream reports in each sample, p a rtly because it is
d iffic u lt to co lle c t large samples o f dreams using the 2-week dream diaries

on w h ich m any researchers tend to rely. In add ition , in tro d u cto ry courses
in statistics emphasize th a t a sample size o f 30 makes it possible to use
sim ple param etric statistics, such as the t test. H ow ever, the use o f sm all
samples in dream research overlooks a m ajor problem : N o t a ll dreams co n ta in
the elements being studied. In actuality, then, the sample sizes in m any
dream studies are even sm aller than they appear to be.
A cco rd in g to C ohen’s (1977, p. 205) detailed w ork on the sample
sizes necessary fo r a tta in in g statistical significance w ith various m agnitudes
o f difference between samples, it takes a large num ber o f observations to
detect sm all differences w ith any degree o f certa inty. For example, w ith a
real difference o f 20 percentage points, w h ich is about as large a difference
as is generally found in dream studies, 125 observations are needed to have
an 80% chance o f a tta in in g statistical significance a t the .05 level. For
differences o f 10 percentage points, it takes a sample o f 502 observations
to have an 80% chance o f a tta in in g significance at the .05 level. Thus,
draw ing one dream from longer dream diaries provided by 30 M exican
Am ericans and com paring them w ith 30 dream reports from A n g lo A m e ri
cans chosen in the same way alm ost guarantees no differences w ill be found
(Kane, M elle n , Patten, & Samano, 1993). S im ila rly, comparisons o f wom en
who w ork at hom e w ith wom en in the paid w ork force are n o t going to
yie ld results th a t can be replicated w hen they are restricted to 30 to 40
dreams per sample by using on ly one dream per dream diary (Lortie-Lussier
et al., 1985).
T H E Q U A L IT Y O F T H E LIT E R A T U R E O N D R E A M C O N T E N T
T his chapter provides a benchm ark fo r assessing the lite ra tu re on dream

content. G enerally speaking, the lite ra tu re is anecdotal, contradictory, and
dispute ridden and is based on c lin ic a l case studies, poorly collected and
overly sm all samples, unreliable and unvalidated ra tin g scales, and inappro
priate sta tistica l m ethods. As the com ments in th is chapter on several studies
suggest, m ost published studies are n o t substantial enough to be used in
developing a neurocognitive m odel. However, rath er th a n p ro vid in g a long
and tedious critiq u e o f the m any weak studies in the lite ra tu re , th is book
proceeds by b u ild in g on the few solid studies th a t exist. For com m ents on
m any o f these inadequate studies, see D o m ho ff (1996, 1999b).
The critiq u e th a t R o th et al. (1979) made o f n in e studies on m edication
and dream co n te n t can be applied to m any o ther parts o f the co n te n t
lite ra tu re . T hey noted th a t “few if any o f the existing studies were m ore
th a n p ilo t studies” and th a t “ there has been a fa ilu re to exam ine even one
drug in depth” (R o th et al., 1979, p. 221). A lth o u g h they concluded th a t
some m edications may affect dream content, they also say th a t “ the la ck
o f standardization o f m ethods o f assessing q u a lity o f dream co n te n t has
resulted in isolated bits o f in fo rm a tio n th a t do n o t yet form a coherent
p ictu re ” (R o th e t al., 1979, p. 221). Kram er and R o th (1979) reached sim ilar
conclusions about the lite ra tu re on dream co n te n t and m ental illness. In a
fo llo w -u p assessment o f the lite ra tu re o n dreams and m ental h e a lth since
1975, Kram er (1999, 2000b) noted th a t the lite ra tu re has n o t im proved.
C O N C L U S IO N
Despite the poor q u a lity o f the lite ra tu re , th is chapter nonetheless

illustrates the p o te n tia l fo r rigorous studies o f dream co n te n t th a t can be
used to test some o f the ideas presented in chapter 1. First, there is every
reason to believe th a t the q u a lity o f dream reports can be h ig h if they are
collected in a sleep laboratory or w ith the m ost recent dream m ethod or if
a lo ng-te rm dream series can be obtained and authenticated. Second, solid
evidence dem onstrates th a t co n te n t analysis, w hen it is based on no m in a l
categories, can be objective and useful. T h ird , good sta tistica l m ethods are
available fo r analyzing the data th a t derive from co n te n t analysis. These
positive conclusions set the stage fo r a detailed presentation o f the H a ll-V a n
de C astle system.

THE HALL-VAN DE CASTLE SYSTEM
T he purpose o f th is chapter is to show th a t the H a ll-V a n de Castle

system o f co n te n t analysis has the necessary re lia b ility and v a lid ity fo r
research th a t lin ks dream con te n t to the neural netw o rk fo r dream ing, on
the one hand, and to w aking cogn ition , on the other. The chapter includes
additions to the system th a t can be used to study h ig h ly m emorable dreams
and describes the statistical methods and new software th a t now make the
system even more accurate and pow erful. T he system is accessible to dream
researchers anywhere in the w o rld because everything needed to carry out
a m odern H a ll-V a n de Castle analysis is available on the In te rn e t at h ttp ://
www.Dream Research.net: coding rules, aids fo r learning the coding system,
a spreadsheet fo r rapid and accurate data analysis called D ream SA T, and
programs fo r using random ization statistics to determ ine p values and co n fi
dence intervals (Schneider & D om hoff, 1995). P rinted copies o f the coding
rules and norm ative findings can be found in H a ll and V an de Castle (1966)
and D o m ho ff (1996).
As b rie fly m entioned in chapter 2, the system rests on the no m in a l
level o f measurement, thereby avoiding the many problem s w ith rating
scales. The o rig in a l eight general categories are as follow s:
■ Characters (subdivided in to anim als, humans, and m yth ica l

figures/creatures)
■ Social Interactions (subdivided in to frie n d ly, aggressive, and
sexual)
■ A c tiv itie s (o fte n analyzed in terms o f physical and nonphysi
cal a ctivitie s)
■ S triv in g (success and fa ilu re )
■ M isfortunes and G ood Fortunes
■ Em otions (anger, apprehension, confusion, happiness, and
sadness)
■ Physical Surroundings (settings and objects)
■ D escriptive Elem ents (m odifiers, tem porality, and n e g a tivity).
Tw o categories have been added: Food and Eating and Elements From
the Past. They were o rig in a lly designed as th eoretical categories to measure
67
“ o ra lity ” and “ regression” (H a ll & V an de Castle, 1966). H ow ever, as noted
in chapter 2, no th e o re tica l categories o f any k in d have proved useful in
the study o f dream content. Fortunately, the descriptive nature o f these tw o
scales makes it possible to adapt them in to em pirical categories. For exam ple,
the m ain categories o f the Food and E ating scale are sim ply (a) eating or
d rin kin g , (b ) being in a restaurant or bar, (c) preparing food, (d) purchasing
or gathering food, and (e) seeing o r m e n tio n in g food. T he Elem ents From
the Past scale consists o f categories such as (a) being younger, (b) seeing a
person w ho has been dead fo r at least a year, and (c) being in a lo ca tio n ,
seeing a person, doing an a c tiv ity , o r seeing an object th a t is m entioned as
n o t being a p a rt o f the dream er’s life fo r at least a year. W ith b o th o f these
scales, it is also possible to elim in a te one o r m ore o f the subcategories fo r
specific research studies.
Extensive experience has shown th a t few aspects o f dream co n te n t do
n o t fit in to one or another o f the 10 H a ll-V a n de Castle categories. T h is
fin d in g , however, does n o t mean th a t a ll the categories need to be used in
every in ve stig a tio n or th a t they have proven to be o f equal value in develop
ing hypotheses and testing theories. T he Big Five o f the coding system are
Characters, S ocial Inte ractio ns, Em otions, M isfortunes, and S triv in g , but
o ther categories have shown themselves to be valuable fo r specific investiga
tions. In the dreams o f Franz Kafka, fo r exam ple, it is the h ig h percentage
o f nonphysical a ctivitie s and the a tte n tio n to the hum an body th a t make
his dreams d is tin c tiv e . M oreover, the differences fit w e ll w ith his w aking
preference fo r observation rather th a n physical a ctivitie s and w ith his obses
sion concerning b o d ily defects (H a ll &. L in d , 1970).
Because the coding categories are clearly defined— the result o f a 2-year
process o f tria l and error in constructing them — the H a ll-V a n de C astle
system has h ig h in tercoder re lia b ility . T h is h ig h re lia b ility is determ ined by
the percentage-of-agreem ent m ethod, w h ic h means th a t a ll the sim ila r cod
ings by tw o independent coders are d ivid e d by the num ber o f agreements
plus the num ber o f disagreements. For exam ple, if coder A makes 51 codings
fo r Characters and coder B makes 49 codings, and they make the same
coding 48 tim es, th e n the intercoder re lia b ility is 48 divide d by 52 (48
agreements plus 4 disagreem ents), w h ic h equals .92. A lth o u g h a s lig h tly
m ore com plex form ula can be used w ith categories w ith e ith e r very h ig h
o r very low frequencies (S m ith , 2000, p. 325), no H a ll-V a n de C astle
categories are at these extremes.
H a ll and V a n de Castle decided to use the percentage-of-agreem ent
approach to com puting re lia b ility by com paring its results w ith w hat they
found w ith every o th e r conceivable m ethod. In fact, they showed th a t the
outcomes from various m ethods o f d e term ining intercoder re lia b ility can
range fro m 0% to 100% w ith the same codings (H a ll & V a n de Castle,
1966, pp. 145 -1 47). I t is therefore meaningless in th e ir eyes to re p o rt a

re lia b ility figure w ith o u t stating the m ethod used to determ ine it, and it
makes little sense to use the other m ethods w ith th is p a rticu la r coding
system (D om hoff, 1996; H a ll, 1969a; V a n de Castle, 1969). M ore generally,
the percentage-of-agreem ent approach is now the one th a t is m ost frequently
used in a ll areas o f co n te n t analysis (S m ith , 2000).
T h e H a ll-V a n de Castle scales fo r various types o f social interactions
avoid the untenable psychological assumptions ofte n im p lic it in ra tin g scales.
Instead o f constructing an o rd in a l aggression scale, fo r example, H a ll and V an
de C astle created eight separate n o m in a l categories fo r types o f aggression:
1. H o stile thoughts
2. C ritic a l remarks
3. Rejections and refusals
4. D ire verbal threats
5. Stealing or destruction o f possessions
6. Being chased
7. Being confined or attacked
8. M urder
Codings can be done reliably, and no in fo rm a tio n is lost w hen eight

d iffe re n t frequencies can be noted and compared. For m any purposes, the
frequencies fo r the eight categories can be com bined to create an aggres
sion score.
A s illu stra te d in chapter 2 w ith the exam ple o f the “ anim al percent,”
findings w ith the no m in a l categories are m ost useful and best understood
w hen they are conveyed in an array o f percentages and ratios th a t are called
content indicators. In a sim ilar fashion, the to ta l num ber o f know n characters,
d ivide d by the to ta l num ber o f a ll characters, leads to the “ friends percent,”
w h ich has proven useful in distinguishing the dreams o f people w ith m ental
illness from norm ative samples (D om hoff, 1996). T he to ta l num ber o f male
characters divide d by the to ta l num ber o f hum an characters whose gender
is indicated yields the “ m ale/fem ale percent,” in w h ic h b o th the m ale and
fem ale percentages are presented to avoid sexism. T he to ta l num ber o f
references to various types o f confusion, unce rta in ty, and surprise, when
d ivid e d by the to ta l num ber o f em otions, provides the “ confusion percent.”
Aggressive and frie n d ly social interactions can be analyzed by determ in
in g rates per character, a m ethod th a t provides an exce lle n t c o n tro l fo r the
num ber o f characters. For example, the to ta l num ber o f frie n d ly interactions
(F ) divide d by the to ta l num ber o f characters (C ) provides a ra tio called
the F /C index. T h is index also can be figured fo r specific characters o r types
o f characters in dreams, leading to an F/C ra tio in dex w ith parents, friends,
or strangers. M athem atically speaking, the ratios d iffe r from the percentage
indicators in the previous paragraph. They therefore require d iffe re n t sta tis ti
cal treatm ent.
T able 3.1 presents a ll the indicators in the H a ll-V a n de C astle system
and explains how they are calculated. T he table includes several indicators
th a t are a sim ple percentage o f the dreams in the sample w ith at least
one instance o f a given category; they are therefore called “ at-least-one”
indicators. T hey are used fo r Aggression, Friendliness, S exuality, Food and
E ating, M isfortune, G ood Fortune, Success, Failure, S triv in g , N egative Em o
tions, and C onfusion. These indicators are useful fo r a q uick com parison o f
tw o samples or fo r studies o f large samples, b u t they have to be used w ith
great ca u tio n because they do n o t include a co rre ctio n fo r the v a ria tio n in
le n g th o f dream reports from sample to sample.
E M O T IO N S IN D R E A M S
T h e coding categories fo r em otions may prove to be p a rticu la rly im p o r

ta n t in lin k in g dream co n te n t to the neural netw o rk fo r dream ing. T he
coding system has five em otions categories: anger, apprehension, confusion,
happiness, and sadness. T he system is lim ite d to these five categories because
it was im possible to develop a reliable one w ith m ore categories. H a ll and
V a n de Castle (1966) w rote th a t “ the classification o f em otions was one o f
our m ost d iffic u lt tasks” (p. 110). C o in cid e n ta lly, good psychological reasons
e xp la in w hy it is hard to o b ta in reliable coding w ith m ore th a n these five
categories. Several d iffe re n t kinds o f psychological studies suggest th a t on ly
a sm all han d fu l o f basic hum an em otions exist. T he m ain five are basic
le ve l categories: anger, fear, love, jo y, and sadness (M u rp h y & Lassaline,
1997; Shaver, Schwarz, K irson, & O ’C onnor, 1987). Some lists also include
contem pt, disgust, and surprise (Ekm an, 1992a, 1992b).
These categories correspond closely to those in the H a ll-V a n de Castle
system. T h e love and jo y categories are in effect collapsed in to one category,
happiness, fo r the sim ple reason th a t the re la tiv e ly few positive em otions
in dreams are d iffic u lt to distinguish in a reliable fashion. A s noted in passing
in chapter 1, o n ly 20% o f a ll em otions in dreams fit in to the happiness
category, a fin d in g th a t has been replicated three tim es (H a ll e t al., 1982;
Roussy, 2000; Tonay, 1990/1991). T h e w aking em otions o f sadness, anger,
and fear (o r apprehension) are exactly represented in the H a ll-V a n de
C astle system. A s fo r disgust and contem pt, they are included w ith in the
anger category.
T he category fo r surprise in the set o f hum an em otions advocated by
Ekm an (1 992b) corresponds w ith the category fo r confusion in the H a ll-V a n
de C astle system. A lth o u g h confusion was n o t considered an em otion at
the tim e the coding system was created, it was included because it is a
psychological state th a t occurs freq uently in dreams. In fact, 12% o f m en’s
dreams and 13% o f w om en’s have at least one instance o f confusion. M ore-

TABLE 3.1
Formulas for Calculating the Hall-Van de Castle Content Indicators
Characters
Male/female percent Males 4 - (Males + Females)
Familiarity percent Familiar 4. (Familiar + Unfamiliar)
Friends percent Friends 4- All humans
Family percent (Family + Relatives) + All humans
Animal percent Animals 4- All characters
Social interaction percents

Aggression/friendliness percent Dreamer-involved aggression + (D-inv. aggression +
D-inv. friendliness)
Befriender percent Dreamer as Befriender 4- (D as Befriender + D as
Befriended)
Aggressor percent Dreamer as Aggressor 4- (D as Aggressor + D as
Victim)
Victimization percent Dreamer as Victim 4- (D as Victim + D as Aggressor)
Physical aggression percent Physical aggressions + All aggressions
Social interaction ratios

Aggression/character index All aggressions + All characters
Friendliness/characters index All friendliness + All characters
Sexuality/characters index All sexuality -5- All characters
Self-concept percents
Self-negativity percent (D as Victim + D-inv. Misfortune + D-inv. Failure) +
(D as Victim + D-inv. Misfortune + D-inv. Failure +
D as Befriended + D-inv. GF + D-inv. Success)
Bodily misfortunes percent Bodily misfortunes 4 - All misfortunes
Negative emotions percent Negative emotions + All emotions
Dreamer-involved success D-involved success 4- (D-inv. success + D-inv. failure)
percent
Torso/anatomy percent Torso, Anatomy, Sex body parts - 5- All body parts
Other indicators
Physical activities percent (P, M, and L activities) 4- All activities
Indoor setting percent Indoor 4- (Indoor + Outdoor)
Familiar setting percent Familiar 4- (Indoor + Outdoor)
Distorted setting percent Distorted settings 4- All settings
Unusual character percent (Dead, imaginary, metamorphoses, and creatures)
All characters
Confusion percent Confusion 4 - All emotions
Percentage of dreams with at
least one
Aggression Dreams with aggression 4 - Number of dreams
Friendliness Dreams with friendliness 4 - Number of dreams
Sexuality Dreams with sexuality 4- Number of dreams
Misfortune Dreams with misfortune 4- Number of dreams
Good fortune Dreams with good fortune 4 - Number of dreams
Success Dreams with success 4- Number of dreams
Failure Dreams with failure 4 - Number of dreams
Striving Dreams with success OR failure 4- Number of
dreams
D = dreamer; inv. = involved; L = location change; M = movement; P = physical; GF = good fortune

over, confusion accounts fo r 22% o f the em otions in m en’s dreams and 18%
o f those in w om en’s dreams, second o n ly to apprehension, w h ich constitutes
45% o f the em otions in m en’s dreams and 37% in w om en’s. C onfusion
in dreams is “ generally produced e ith e r through c o n fro n ta tio n w ith some
unexpected event or else through in a b ility to choose between available
alternatives” and is in dica ted by such term s as “ surprised, astonished, amazed,
awestruck, m ystified, puzzled, perplexed, strange, bew ildered, dou btful, con
flicte d , undecided, and unce rtain” (H a ll & V a n de Castle, 1966, p. 112).
I t is lik e ly th a t em otion know ledge is organized in at least tw o ways:
by a p o sitive -n e g a tive dim ension and by in te n s ity (Shaver et al., 1987).
B o th o f these dim ensions are reflected in the H a ll-V a n de C astle system.
T h e p o sitive -n e g a tive dim ension is captured by th e in clu sio n o f a “ negative
em otions percent” as one o f the co n te n t indicators in T able 3.1. A s fo r the
in te n s ity o f em otions, it can be coded w ith the H a ll-V a n de Castle system’s
In te n s ity scale fo r classifying any w ord m odifiers th a t are used to describe
force or the expenditure o f energy. These m odifiers may refer to e ith e r
physical or m ental energy or to em otions and sensations. In the case o f
em otions, the sim ple appearance o f the em otion is n o t sufficien t to code
fo r in te n sity. T he dreamer must use some in te n s ify .m odifier, such as “ very”
or “ extrem ely,” fo r the em otion to be coded 1+ (h ig h in te n s ity ), or “ m ild ly ”
or “ a little ” fo r it to be coded I - (lo w in te n s ity ).
N O R M A T IV E F IN D IN G S O N C O LLE G E M E N A N D W O M E N
T o make the system m ore useful fo r studying in d ivid u a ls o r unique

p o p u la tio n groups, H a ll and V an de C astle (1966, chapter 14) created a
set o f norm ative findings th a t were based o n five dreams fro m each o f 100
m ale and 100 fem ale students at Case W estern Reserve U n iv e rs ity and
B aldw in-W allace C ollege in C leveland, O h io , between 1947 and 1950.
T able 3.2 provides the norm ative findings fo r m ost o f the co n te n t indicators
presented in T able 3.1. O th e r norm ative findings are available upon request
through http://w w w .D ream R esearch.net. These results, in c lu d in g a b rie f
discussion o f the gender sim ilarities and differences, are presented la te r in
th is chapter.
T he H a ll-V a n de Castle norms were replicated fo r m en and wom en
at the U n iv e rs ity o f R ichm ond in the early 1980s; fo r w om en at the U n iv e r
sity o f C a lifo rn ia , Berkeley, in the m id-1980s; a t Salem C ollege in the late
1980s; and at the U n iv e rs ity o f C a lifo rn ia , Santa Cruz, in the early 1990s
(D om hoff, 1996; D udley & Fungaroli, 1987; D udley & Swank, 1990; H a ll
et al., 1982; Tonay, 1990/1991). T he norm s fo r several o f the character
categories were replicated at the U n iv e rs ity o f C in c in n a ti in the late 1960s

TABLE 3.2
A Comparison of the Male and Female Norms for the
Hall-Van de Castle System
M ale norms Fem ale norm s Effect
(%) (%) size P
Characters
M ale/fem ale percent 67 48 +.39 .000**
Fam iliarity percent 45 58 - .2 6 .000**
Friends percent 31 37 - .1 2 .004**
Fam ily percent 12 19 -.2 1 .000**
A nim al percent 6 4 +.08 .037*

A ggression/friendliness percent 59 51 +.15 .014*
B efriender percent 50 47 +.06 .517
A ggressor percent 40 33 +.14 .129
Victim ization percent 60 67 - .1 4 .129
Physical aggression percent 50 34 +.33 .000**

A ggression/character index .34 .24 +.24 .000**
Friendliness/character index .21 .22 -.0 1 .852
S exuality/character index? n . .06 .01 +.11 .000**
Self-concept percents
S elf-negativity percent 65 66 - .0 2 .617
Bodily m isfortunes percent 29 35 - .1 2 .217
N egative em otions percent 80 80 +.00 .995
D ream er-involved success percent 51 42 +.18 .213
T orso/anato m y percent 31 20 +.26 .002**
Other indicators
Physical activities percent 60 52 - .3 8 .000**
Indoor setting percent 48 61 - .2 6 .000**
F am iliar setting percent 62 79 - .3 8 .000**
Percentage of dreams with

at least one
A ggression 47 44 +.05 .409
Friendliness 38 42 -.0 8 .197
S exuality 12 4 +.31 .000**
M isfortune 36 33 +.06 .353
Good fortune 6 6 +.02 .787
Success 15 8 +.24 .000**
Failure 15 10 +.17 .007**
Striving 27 15 +.31 .000**
A/ofe. The p values are based on the formula for the significance of differences between two proportions.
The effect size derives from Cohen’s h. The h statistic is determined by the following formula:
h = c o s 1(1 -2 P 1)-co s-1(1 ~2P2)
P, and P2 are proportions between 0 and 1, and the cos ’ operation returns a value in radians.
* significant at the .05 level
"sig n ifica n t at the .01 level
73
(Reichers, Kram er, & T rin d e r, 1970). T he seeming exceptions to these
replication s (B ursik, 1998; Kram er et ah, 1983; R ubenstein & K rip pner,
1991) have m ajor m ethodological problem s th a t are discussed in d e ta il
elsewhere (D om hoff, 1996, 1999b).
T he norms o rig in a lly were calculated by p o o lin g a ll 500 dreams in
each sample, thereby ig noring the fa ct th a t they are n o t independent obser
vations. T he same results, however, are obtained w hen the figures fo r each
o f the 100 particip ants in each sample are com puted and th e n averaged
across individuals. Some o f the evidence fo r th is p o in t is presented in Table
3.3. T h is recalcu latio n answers the concern expressed by U rb in a (1981)
th a t determ ining the results fo r each person and then averaging them m ig h t
lead to d iffe re n t results.
T he H a ll-V a n de Castle system sometimes can be made m ore useful
by com bin ing tw o o r m ore o f its no m in a l em pirical categories. Such com bina
tio n s can be used instead o f relying on th e o re tica l ra tin g scales o f unkn ow n
re lia b ility and v a lid ity to test new hypotheses. T h is p o in t can be dem on
strated firs t o f a ll by lo o kin g at Beck and H u rv ic h ’s (1959) m islabeled and
unvalidated M asochism scale, w h ich is one o f the few scales o ther th a n the
H a ll-V a n de Castle system th a t has been used by independent investigators.
T h e scale consists o f a w ide range o f negative experiences ranging from
physical discom fort to re je c tio n to being punished, lost, o r victim ize d. U sing
th is scale, C a rtw rig h t (1992) came to the conclusion th a t divorced wom en
w ho are n o t depressed are more m asochistic th a n divorced m en w ho are
depressed. T h is is a surprising result th a t seems to raise m ore questions th a n
it answers.
TABLE 3.3
Group vs. Individual Norm Comparisons for Selected Categories
C ollective A verage o f 100
m ale norm s individual norm ative D ifference
(%) dream ers (%) (%)
M ale/fem ale percent 67.8 66.1 - 1 .7
F am iliarity percent 45.0 44.9 -0 .1
Friends percent 31.4 30.8 - 0 .6
A nim al percent 6.0 5.9 -0 .1
A g g ression /frie ndliness percent 59.0 58.6 - 0 .4
A ggre ssio n /ch a ra cte r index .340 .371 .031
F riendliness/ch aracter index .211 .210 -.0 0 1
Indoor setting percent 48.6 49.5 0.9
S elf-nega tivity percent 64.7 66.0 1.3
A t least one A ggression 47.2 47.2 0.0
A t least one Friendliness 38.2 38.2 0.0
A t least one S exuality 11.6 11.6 0.0
A t least one M isfortune 36.4 36.4 0.0

In a useful study, C lark, T rin d e r, Kram er, R o th, and Day (1972) h y
pothesized th a t the item s on the M asochism scale are encompassed by three
categories in the H a ll-V a n de Castle system: failures, m isfortunes, and
v ic tim status in aggressive interactions. They th e n dem onstrate th is p o in t
by coding tw o d iffe re n t samples w ith b o th the M asochism scale and the three
H a ll—V a n de Castle categories. They found th a t the masochism findings were
encompassed by the H a ll-V a n de Castle categories, w h ich also located
several elements missed by the M asochism scale.
Because wom en are s lig h tly m ore lik e ly to fa il w hen they strive in
dreams and to be victim s in aggressive interactions, the greater masochism
th a t C a rtw rig h t reported in her women participants was really a com bination
o f failures and victim iza tions, w h ich are n o t m anifestations o f c lin ic a l mas
ochism . M oreover, the H a ll-V a n de Castle m isfortune categories, w h ich
m ig h t seem to be related to masochism, do n o t show any gender differences.
T he three categories are part o f the “self-neg ativity percent” explained in
Table 3.1. T h is category, along w ith a lack o f friends and frie n d ly in te ra c
tions, distinguishes the dreams o f people w ith m ental illness from those o f
people w ith o u t m ental illness (D om hoff, 1996, 1999b, p. 127; M aharaj,
1997).
T he usefulness o f com bining H a ll-V a n de C astle categories to encom
pass speculative th eoretical ra tin g scales also can be shown by lo o kin g at
K ro h n and M aym an’s (1974) com plicated system fo r determ ining the le vel
o f m a tu rity in object relations, a concept draw n from a B ritish va ria n t o f
classical psychoanalytic theory. T he scale calls fo r subtle judgm ents such
as assigning an 8 if the dreamer reports “ a sense o f rapport w ith people and
a w ell-developed understanding o f th e ir thoughts, feelings and co n flicts”
(K ro h n & M aym an, 1974, p. 454). A 5 is assigned if the people in the
dream have “no real id e n tity ” ; a 3 if people are experienced as “ insubstantial,
flu id , m ore or less interchangeable” ; and a 1 “ i f there are no o ther people
and the subject’s w o rld seems to be com pletely lifeless, vacant, alien, strange”
(K ro h n & M aym an, 1974, pp. 4 5 2 -4 5 4 ). W inegar and L e vin (1997) applied
the scale, w h ich never has been validated, to 389 dream reports recorded
by 115 adolescents between ages 15 and 18. T hey found th a t the girls showed
m ore m a tu rity in object relations th a n did boys and th a t the differences were
greater at the older ages.
W inegar and L e vin provided copies o f the dream reports used in th e ir
study fo r reanalysis fo r th is book. A sample o f 12 reports coded “ lo w ”
and 10 reports coded “ h ig h ” on the K ro h n -M a ym a n scale was coded w ith
H a ll-V a n de Castle categories by tw o coders w ho had no knowledge o f the
K ro h n -M a ym a n scale or o f the fa ct th a t tw o sets o f dreams were being
used. They found th a t three easily coded no m in a l categories— frie n d ly in te r
actions, activitie s, and physical aggression— distinguished the tw o sets o f
dreams alm ost perfectly. Dreams w ith at least one frie n d ly in te ra ctio n were
always in the h ig h group, as were dreams w ith nonphysical a c tiv itie s lik e
ta lkin g . Dreams w ith physical aggressions and physical a ctivitie s were in
the low group w ith one exception, in w h ich a frie n d ly in te ra c tio n also to o k
place. In effect, these results reveal the basis on w h ich raters are m aking
th e ir judgm ents. T hey also suggest th a t the H a ll-V a n de Castle norm ative
findings on physical a ctivitie s percent, physical aggression percent, aggres-
sion/friendliness (A /F ) percent, and the percentage o f dreams w ith at least
one friendliness could be used re lia b ly to assess the re la tive m a tu rity or
im m a tu rity o f a group o r in d ivid u a l.
PR ESER VIN G SEQ U EN CES A N D F IN D IN G C O N N E C T IO N S
T he H a ll—V a n de Castle system contains tw o general categories fo r

dealing w ith sequences o f events in dreams. In the case o f social in teractio ns,
it is possible to code fo r reciprocated aggressive, frie n d ly, or sexual in te ra c '
tions, such as a character sm ilin g back after being extended a frie n d ly
greeting. In the case o f m isfortunes, good fortunes, successes, and failures,
it is possible to code fo r consequences. For example, a fa ilu re can be follow ed
by a good fo rtu n e th a t reverses the failure , a new e ffo rt th a t leads to success,
or a better outcom e as the result o f the frie n d ly in te rv e n tio n o f another
dream character. T he m ost in terestin g fin d in g from sequential analyses is
how rarely they occur. T here are very few “ reciprocities” fo r e ith e r aggressive
or frie n d ly in teractio ns, and even fewer “ consequences” o f m isfortunes, good
fortunes, successes, and failures. H ow ever, these categories may prove to be
useful in the study o f in d iv id u a l dream journals.
Patterns and relationships among coding categories can be studied by
means o f contingency analysis, a nonparam etric sta tistic designed fo r use
w ith n o m in a l data th a t provides the exact p roba bilities o f relationships
between tw o elements. I t is id eally suited fo r co n te n t analysis studies (e.g.,
Osgood, 1959). C ontingencies fo r any p a ir o f H a ll-V a n de C astle coding
categories can be found using the SearchCodings program available through
Dream Research.net. O nce codings have been entered, the categories fo r
w h ich contingency in fo rm a tio n is sought are entered in designated in p u t
boxes. T h e program calculates th e observed and expected frequencies o f
the overlaps and provides the p value fo r the difference th a t is found. For
example, if th e coding categories fo r strangers and physical aggressions are
selected and b o th the m ale and fem ale norm ative samples are searched, the
results reveal a sig n ifica n t contingency between m ale strangers and physical
aggressions fo r b o th m en and wom en, b u t the contingency does n o t h o ld
for fem ale strangers. These results are displayed in T able 3.4.

TABLE 3.4
Contingency Between Physical Aggression and Strange Male Adult
Characters (xMSA) in the Male and Female Norms
C ontingency between C ontingency betw een
physical aggression and physical aggression and
m ale strangers fem ale strangers
O bserved Expected O bserved Expected
(%) (%) P (%) (%) p
M ale norms 0.324
Fem ale norm s 0.395
M E M O R A B LE D R E A M S W IT H A N IM P A C T O N T H E D R E A M E R
M ost studies using the H a ll-V a n de Castle system have focused on

everyday dream reports. However, the system can be adapted to study the
occasional h ig h ly m emorable dreams th a t have carryover effects in to w aking
life and are o f great interest to clin icia n s and hum anists (B ulkeley, 1999;
Knudson &. M in ie r, 1999; K uiken & Sikora, 1993). T hen, too, the H a ll-V a n
de Castle norms provide a basis fo r com m enting on the frequency o f
such dreams.
M em orable dreams— dreams th a t have an im pact on the dreamer—
are often characterized by w hat H a ll and V an de Castle ca ll “ good fortunes,”
w h ich are defined as good things th a t happen to a dream character as a
result o f fate, rather th a n as a result o f successful s triv in g by the character
or frie n d ly actions by other dream characters:
A good fortune is coded when there is an acquisition o f goods or
something beneficial happens to a character that is completely adventi-
tious or the result of a circumstance over which no one has control.
A good fortune is also scored if the dreamer appears in a bountiful
environment. (H all & Van de Castle, 1966, p. 105)
H a ll and V a n de Castle contrast good fortunes w ith m isfortunes, w h ich

are bad outcomes th a t occur to characters independent o f any aggressive
interactions or failures. They found th a t m isfortunes occur in 36% o f m en’s
dreams and 33% o f wom en’s dreams, whereas good fortunes appear in on ly
6% o f dreams fo r b o th men and women. Because o f the sm all num ber o f
good fortunes, they created o n ly one general category fo r good fortunes. For
purposes o f studying memorable dreams, however, it is necessary to d ivide
th is category in to six categories th a t parallel the six categories fo r m isfortunes
(B ulkeley, D unn, & D om hoff, 2001).
T he six expanded good fortune (G F ) categories are defined as the
m irro r opposites o f the six M isfortunes (M ) categories, so it is useful to
THE HALL-VAN DE CASTLE SYSTEM 77

e xp la in them in term s o f th a t contrast. Because M 6 is defined as “ a character
is dead or dies as a result o f accident o r illness or some unknow n cause,”
G F6 is defined as a “ character returns from the dead, acts as if he or she
were alive, or transform s from an inanim ate character (such as a puppet)
in to an anim ate character.”
Just as M 5 encompasses characters w ho are suffering in ju ry , illness,
pain, m edical operations, b o d ily o r m ental defects, insanity, amnesia, or
blindness, so G F5 encompasses recovery from illness or in ju ry, m iraculous
healing, possession o f extraordinary b o d ily powers, sw im m ing under w ater fo r
long periods w ith o u t any need to breathe, enhanced m ental and perceptual
a b ilitie s, extrasensory perception, and dream lu c id ity . (T he a b ility to fly
under one’s ow n pow er is n o t coded as a GF5 because it has its ow n category,
G F2, to contrast w ith M 2 , w h ich is fo r fa llin g or the danger o f fa llin g .)
T he category M 4 focuses p rim a rily on any character’s lost, damaged,
o r defective possessions, b u t it also includes accidents w ith o u t in ju rie s th a t
are experienced by any character. A lost rin g , a house destroyed by a fa llin g
boulder, a broken com puter, and a car crash because o f an icy road are a ll
coded as M 4. C onversely, GF4 is defined by the possession o f enhanced,
m agical, o r supernatural objects and by experiencing m iraculous events th a t
are, in effect, th e opposites o f car crashes on icy roads, such as cars th a t fly
over the traffic.
T he category M 3 includes a ll threats to a character’s w e ll-b e in g and
em otion al s ta b ility th a t result from an im personal force in the enviro nm ent,
such as walls or trees about to fa ll on the dreamer, boats about to capsize,
hurricanes th a t m ig h t in ju re the dreamer, and tid a l waves th a t m ig h t eng ulf
the dreamer. GF3 therefore includes a ll situations in w h ich an im personal
force in the e n viro n m e n t may enhance a character’s w e ll-be ing o r h e ig h te n
em otion al pleasure or satisfaction, such as waves th a t carry a person to the
shore or b e a utifu l rainbow s th a t cause great joy.
M 2, as already noted, is reserved fo r fa llin g or the danger o f fa llin g .
For sym m etry’s sake, GF2 is restricted to any situ a tio n in w h ic h a dream
character is suspended above the ground o r flyin g under her o r his ow n
power. T h is category also can include situations in w h ich a dream character
feels as though she or he has the p o te n tia l to fly, liv e in the clouds, be in
another universe, o r be in any other state th a t is the opposite o f “ the danger
o f fa llin g .”
F in a lly, M l encompasses a ll those fru stra tin g situations in w h ic h dream
characters encounter enviro nm ental obstacles, such as doors th a t w on’t
open, or in w h ich they are unable to m ove, are lost, or are in danger
o f being late. GF1 is therefore defined by the “ b o u n tifu l environm ents”
m entioned in H a ll and V a n de C astle’s (1966) d e fin itio n o f a good fortune .
These are environm ents w ith bea utifu l anim als, im aginary creatures, bea uti
fu l vistas, and an abundance o f resources. GF1 differs from GF3 in th a t the

dream report its e lf contains no evidence th a t the dream character’s w e ll
being is heightened or em otional satisfaction enhanced.
In a d d itio n to the frequent presence o f good fortunes, m emorable
dreams also seem to be characterized by p a rticular patterns o f H a ll—V a n de
Castle indicators. T h is can be seen by lo o kin g at the patterns in the three
types o f such dreams— transcendent, e xisten tial, and anxiety— id e n tifie d in
studies using 1-m o n th diaries from college students; these three types were
th e n compared w ith everyday dreams (B usink Sr K uiken, 1996; K u iken &
Sikora, 1993). T he examples provided in these tw o studies suggest th a t the
three types o f m emorable dreams f it the fo llo w in g coding patterns:
■ Transcendent dreams are characterized by unusual or distorted

settings, famous or m yth ica l characters, frie n d ly interactions,
good fortunes, success, and positive em otions.
■ Existential dreams are more lik e ly to in vo lve deceased characters,
m ajor m isfortune, th in k in g instead o f physical a c tiv ity , and
intense feelings o f sadness o r confusion.
■ Anxiety dreams co n ta in u n fa m ilia r settings, unknow n characters
or anim als, physical aggression directed at the dreamer, physical
a ctivitie s instead o f th in k in g , and intense feelings o f appre
hension.
■ O rdinary dreams are far m ore lik e ly to have fa m ilia r settings,
fa m ilia r characters, nonphysical aggressions, m in o r m isfortunes,
and no em otions.
C O R R E C T IN G FO R V A R IA T IO N S IN REPO RT L E N G T H •
As discussed in chapter 2, w ide variations in the le n g th o f dream

reports lead to seem ingly contradictory findings in the lite ra tu re on dream
content. Thus, a good co n tro l fo r the d iffe rin g lengths o f dream reports is
essential. In th e ir o rig in a l study, H a ll and V a n de Castle (1966) dealt w ith
th is problem by lim itin g th e ir norm ative sample to dream reports o f between
50 and 300 words.1 However, experience since they w rote suggests th a t the
percentages and ratios discussed earlier in the chapter are a better way to
correct fo r dream le ngth. These in dica tors elim ina te the problem th a t some
people are w ordier th a n others and allow researchers to use dreams o f more
th a n 300 words.
T he spreadsheet on Dream Research.net makes it possible to conduct
large-scale studies to dem onstrate the power o f percentages and ratios in
U n fo rtu n a te ly , subsequent w ork w ith the o rig in a l dream reports and coding cards showed th a t four
male dream reports and one female dream re p o rt in the sample had fewer tha n 50 words, and one
male report and three female reports had m ore tha n 300 words.

corre cting fo r report le ngth. T hree d iffe re n t studies— one using the male
norm s, another the fem ale norm s, and a fin a l one using the “ Barb Sanders”
series (analyzed at le n g th in chapter 5 )— show th a t the con te n t indicators
are successful in e lim in a tin g any biases created by report length. A s a startin g
p o in t, Table 3.5 presents the results o f a co rre la tio n a l study o f w ord le n g th
and raw frequencies. For m ost co n te n t categories, the correlations are low .
A s m ig h t be expected, how ever, there is a substantial co rre la tio n between
the num ber o f words in the dream reports and the num ber o f a ctivitie s,
objects, and characters. T here is also a co rre la tio n between the num ber o f
words and num ber o f aggressions (.25) and between num ber o f words and
the num ber o f frie n d ly in te ra ctio n s (.2 1 ). A lth o u g h n o t shown in the table,
a co rre la tio n o f .32 has been found between characters and aggressions and
.31 between characters and frie n d ly interactions.
Figure 3.1 shows th a t the co n te n t indicators c o n tro l fo r these correla
tions. C om paring dreams w ith fewer th a n 175 words w ith those w ith more
th a n 175 words in three d iffe re n t samples th a t range from 50 to 300 words
reveals no systematic differences between short and long dreams on any o f
the percentages and ratios w hen the tw o samples are compared. T he physical
aggression percent fo r wom en is unexpectedly low er in the long dreams,
and the fa m ilia r setting percent low er in the m en’s dreams, but the same
differences are n o t found in the o th e r samples. S im ila rly, the b o d ily m isfor
tunes percent is low er fo r Barb Sanders in the long dreams, but the same
difference does n o t show up w ith the o th e r tw o samples. O n the o th e r hand,
and as expected, the long dreams are m ore lik e ly to have at least one
in c id e n t o f m ost elements, dem onstrating th a t at-least-one indicators— and,
by inference, the ra tin g scales used in m ost dream studies— are extrem ely
sensitive to dream length.
T h e largest study o f the issue o f dream le n g th is based on the Barb
Sanders series. I t includes dreams th a t range in le n g th fro m 2 to 500 words.
T h is range encompasses a little more than 95% o f her dream reports. T he
sta rtin g p o in t was the random sample o f 250 dream reports th a t ranged
fro m 50 to 250 words in le ngth. T he dreams were coded fo r Characters,
S ocial Inte ractio ns, M isfortunes, and Em otions. T h e n random samples o f
200 short and 200 long dream reports were added; these samples were on ly
coded fo r S ocial Inte ractio ns, the m ost im p o rta n t and sensitive coding
categories, because o f lim ite d resources.
T h e study reveals th a t the at-least-one indicators rise consistently as
w ord le n g th increases, reaching a plateau at 400 words. T h is fin d in g shows
th a t lo ng reports have a greater num ber o f social interactions, as m ost
researchers m ig h t expect. I t is also lik e ly th a t m ost o ther coding categories
w ould show the same se n sitivity to report length. N one o f the rates and
percentages, how ever, is affected by dream le n g th once a m inim u m o f 50
words is reached. A selection o f these results is displayed in Figure 3.2. A t

TABLE 3.5
Correlation of Category Frequencies With Word Length for Males, Females, and Barb Sanders
M ale norm s Fem ale norm s Barb Sanders
/ C orrelation with f C orrelation w ith f C orrelation w ith

per dream w ord count per dream w ord count p e r dream w ord count
O bjects 4.89 +.57 5.33 +.58 n/a n/a
Activities 4.78 +.59 4.95 +.60 n/a n/a
C haracters 2.36 +.46 2.88 +.45 3.68 +.64
M odifiers 2.26 +.39 2.89 +.43 n/a n/a
Settings 1.27 +.24 1.29 +.24 n/a n/a
A ggressions 0.80 +.25 0.67 +.20 1.20 +.18
E m otions 0.57 +.33 0.84 +.27 0.98 +.37
F riendliness 0.50 +.21 0.61 +.27 1.18 +.32
M isfortune 0.41 +.07 0.41 +.07 0.65 +.31
Failure 0.17 +.15 0.10 -.0 1 0.05 - .0 4
Success 0.16 +.17 0.07 +.08 0.06 +.06
S exuality 0.15 +.01 0.04 -.0 1 0.34 +.05
G ood fortune 0.07 +.09 0.06 - .0 6 0.02 +.03
THE HALL-VAN DE CASTLE SYSTEM

00
Figure 3.1. A com parison of long and short dream s from various dream series, with
the shorter dream s serving as the baseline. T he differences in this figure are e xpressed
in term s o f th e effect size statistic h, w hich is a sim ple transform ation o f percentage
differences. An h of .20 o r less is considered sm all. For instance, the fig u re show s
th a t the percentage of settings in the longer dream s that are fa m ilia r to th e m ale
dream ers is m uch low er than in the shorter dream s th a t are serving as the baseline,
which is indicated by the vertical line in the m iddle of the figure. C onversely, th e figure
show s th a t m ore of the longer dream s have at least one aggression and at least
one friendliness than th e sh o rte r dream s fo r m ale dream ers, fem ale dream ers, and
Barb Sanders.

h vs. Baseline 250
-1.0 -0.8 -0.6 -0 .4 -0.2 0 +0.2 +0.4 +0.6 +0.8 +1.0
-.2 0
Aggression/Friendliness %
I +.06
+.06
Befriender %
□ Barb Sanders, 25-49 Words
■ Barb Sanders, 300-499 words
Aggressor %
Physical Aggression %
-.67
A t Least One Aggression
A t Least One Friendliness
A t Least One Sexuality
Figure 3.2. A co m p a riso n of Barb S a n d e rs ’s long and sh o rt dream reports, com pared
w ith h e r dream reports containing 5 0 to 300 w ords. T he differences in this figure are
exp re sse d in te rm s o f th e effect size s ta tis tic h. For instance, this figure show s that
th e aggressor p e rce n t is 41 h points lo w e r in the shorter dream s than it is in the
b a se lin e sam ple of dream s with 50 to 3 0 0 words. C onversely, it show s th a t the longer
d re a m s are 76 h points higher than th e baseline sam ple of dream s w ith at least
o n e friendliness.
th e same tim e, the results in Figure 3.2 also clearly dem onstrate th a t the
co n te n t indicators do n o t do a good jo b o f correcting fo r the distortions
th a t appear in dream reports o f 25 to 49 words. T he findings fo r dreams
under 25 words in length, w hich are n o t included in Figure 3.2, are even

more distorted. T h is result reinforces the earlier decision by H a ll and V a n
de Castle to exclude dream reports shorter th a n 50 words. I t also fits w ith
Hobson, Pace-Schott, and S tickg o ld ’s (2000a, p. 1024) im pressionistic con
clusion th a t it takes at least 50 words to describe a dream experience. Thus,
it seems u n lik e ly th a t repeatable and s cie n tifica lly useful results can be
obtained w ith short reports.
T H E S T A T IS T IC A L A N A L Y S IS O F
H A L L -V A N DE C A S T L E D A T A
Testing fo r S tatistical Significance
M any psychologists believe th a t the usefulness o f determ ining p values

has been vastly overrated because statistical significance is sim ply a fu n c tio n
of sample size if even a sm all difference exists (C ohen, 1994; M eehl, 1997;
Thompson, 1999a). Significance levels may be so m isleading th a t some
psychologists have advocated th e ir banishm ent fro m scie n tific journals
(Hunter, 1997; Scarr, 1997; S chm idt, 1996). Nevertheless, p values are
likely to persist fo r some tim e; studies o f the statistics used in jo u rn a l articles
in the 1990s suggest th a t the critics have had little im pact in deemphasizing
significance levels and increasing the use o f effect sizes (Thom pson,
1599b, 1999c).
T he sta tistica l approach most com patible w ith the H a ll-V a n de Castle
content indicators is sim ple b u t pow erful and uses th e latest techniques
available fo r determ ining p values. A H a ll-V a n de Castle analysis begins
with nonparam etric statistics because they do n o t require any assumptions
abut th e le vel o f measurem ent or the shape o f the sam pling d is trib u tio n .
Although evidence indicates th a t param etric tests can be useful even w hen
some o f th e ir und erlyin g assumptions are vio la te d , o th e r evidence suggests
that nonparam etric tests are better w hen sample sizes are sm all or unequal
(Sawilowsky & B lair, 1992). M oreover, param etric statistics are especially
weak w ith o rd in a l or n o m in a l data, even w hen sample sizes are re la tiv e ly
large, because such data are lik e ly to deviate from a norm al curve (M ic c e ri,
1989; N anna & Saw ilow sky, 1998).
G ive n the la ck o f an underlying continuous d is trib u tio n w ith n o m in a l
data such as the H a ll—V a n de Castle system provides, m ost statistics te x t
books suggest the use o f proportions or ch i square (e.g., Siegel & C astellan,
1988). Fortunately, these tw o statistics are ideal fo r the percentage indicators
used in th e H a ll-V a n de C astle system (e.g., Reynolds, 1984). M oreover,
The test fo r the significance o f differences between tw o independent propo r
tions and c h i square provide exactly the same results fo r th e 2 x 2 categorical
tables most freq uently used w ith the H a ll-V a n de C astle system. T h a t is,

the z score derived from a proportions test is equal to the square ro o t o f
ch i square (Ferguson, 1981, pp. 2 1 1 -2 1 3 ). In add ition , as C ohen (1977)
explains, a p ro p o rtio n is merely a type o f mean in w h ich a ll the values in
the d is trib u tio n are e ith e r 0 or 1, so “ the same k in d o f in fe re n tia l issues”
are in vo lve d w ith proportions as w ith means in general (p. 179). N o th in g
w ould be gained by determ ining th e m ean num ber o f characters o r em otions
or aggressions per dream, even i f th e ca lcula tion o f means d id n o t have
problem s stem m ing from the d iffe rin g lengths o f dream reports from sample
to sample.
T he co m putation ally intensive random ization strategies made feasible
by programs available through Dream Research.net are used to determ ine p
values (N oreen, 1989). L ike the nonparam etric statistics to w h ich they are
closely related, random ization strategies have the great v irtu e o f bypassing
m ost o f the assumptions necessary fo r the use o f param etric statistics. First,
they make random sam pling unnecessary. Second, they do n o t assume a
norm al d istrib u tio n , T h ird , random ization techniques are also useful because
they w ork equally w e ll w ith lo n g itu d in a l studies o f in d iv id u a l cases and
o ther types o f repeated-measures designs. Fourth, thro ugh the use o f 1,000
o r more resamplings fro m the o rig in a l samples, they provide exact p values,
n o t approxim ations. T h e y do assume th a t the samples being compared have
sim ila r shapes and variances, b u t th a t is also true fo r param etric statistics
(F ra n klin , A llis o n , & Gorman, 1997).
T he p values for the percentage differences between samples are deter
m ined by th e random ization strategy called permutation testing by statisticians
and approximate randomization b y social scientists. W ith th is m ethod, the
difference between the tw o o rig in a l samples is com pared w ith a d is trib u tio n
o f differences obtained by pooling the data from b o th samples and th e n
creating 1,000 o r more pairs o f random samples. T h e p value is sim ply the
percentage o f tim es th a t the difference between a pair o f random ly draw n
samples is greater than o r equal to th e difference between the tw o o rig in a l
samples. For example, in a com parison o f the norm ative m ale/fem ale percents
fo r m en (67/33) and women (4 8 /5 2 ), there were o n ly 6 instances in 5,000
tria ls where the difference between the random ly drawn pairs o f samples
was greater th a n or equal to the difference between the tw o o rig in a l samples,
so the p value is .0012 (6/5000 = .0012).
W hen the results using approxim ate random ization are compared w ith
th e standard form ula in a study o f m ale/fem ale percent, there is no disagree
m ent as to the level o f significance. However, w hen the same analysis is
done w ith some o f the other H a ll—V a n de Castle indicators, discrepancies
between the tw o methods are revealed. T he standard form ula sometimes
indicates th a t the results are s ig n ifica n t at the .05 or .01 le vel o f confidence,
b u t approxim ate random ization shows th a t the results are n o t sig nifica nt at
e ith e r o f those levels. In such cases, the form ula is wrong because the

frequency d is trib u tio n is far d iffe re n t from the assumed norm al d is trib u tio n .
Figure 3.3 presents examples o f the h ig h ly skewed frequency d istrib u tio n s
fo r several H a ll-V a n de Castle co n te n t categories. I t also shows th a t the
d is trib u tio n fo r characters is considerably less skewed.
S ta tistica l analyses fo r the aggression (A /C ) and friendliness (F/C )
indexes require a d iffe re n t rationa le because they are ratios, n o t proportions.
S ta tis tic a lly speaking, the S ocial In te ra ctio n indexes are ratios o f tw o sample
means. In the case o f the A /C index, fo r exam ple, the in dex in effect
compares the mean num ber o f aggressions in a sample to the mean num ber
o f characters. Because b o th o f these means are determ ined by adding up
raw frequencies and d iv id in g by the to ta l num ber o f dreams, it is possible
to add up the num ber o f aggressions and d ivid e by the num ber o f characters
to arrive at the same result. O nce the ratios have been calculated fo r b o th
samples and the difference between the tw o samples has been determ ined,
the p value can be determ ined by approxim ate random ization in the same
way as fo r percentage-based indicators. In o th e r words, fo r significance testing
the social in te ra c tio n indexes “ behave” as though they were proportions,
even though they are n o t.
Figure 3 3 F requency distributions fo r selected H a ll-V a n de C astle categories.

D eterm ining Confidence Intervals by Bootstrapping
Confidence intervals provide an in d ic a tio n o f the precision o f the

findings w ith any given sample. C ohen (1994) and M eehl (1997) believed
it w ould be m ore useful— and m ore sobering— if 95% confidence intervals
were provided, rather th a n p values. C ohen (1994) suspects th a t confidence
intervals are seldom reported because they are “ embarrassingly large” in most
psychological studies (p. 1002). T h is p o in t holds true fo r some H a ll-V a n de
C astle findings as w ell. In the case o f the difference between m en and
w om en on the fa m ilia rity percent, fo r example, a 95% confidence in te rva l
o f 18% to 36% does n o t sound as impressive as a p value o f .000 (D om hoff,
1996, p. 56).
The random ization strategy called bootstrapping makes it possible to
determ ine confidence intervals fo r in d iv id u a l samples. A s w ith o ther ran
dom ization statistics, the basic idea is to draw 1,000 o r m ore resamples and
exam ine the d is trib u tio n o f outcomes. In the case o f bootstrapping, the
unique aspect is th a t random samples o f the same size as the o rig in a l sample
are created by p u ttin g the score just draw n back in to the sample before the
n e xt random p ick is made (i.e., “ resam pling w ith replacem ent” ). A s a result
o f replacem ent, some o f the o rig in a l codings appear more th a n once in
each new random sample, and some codings appear n o t at a ll. The basic
assum ption is th a t the resamples “ are analogous to a series o f independent
random samples” (M ooney & D uval, 1993, p. 11). S im u la tio n studies using
samples w ith know n parameters support this assum ption. T hey also show
th a t bootstrapping provides m ore accurate confidence intervals th a n do
standard form ulas th a t assume norm al distrib u tio n s based on random ly drawn
samples. Bootstrapping also allows confidence intervals to be asym m etrical
w hen a d is trib u tio n is skewed, som ething n o t possible w ith the standard
form ula.
T he approach to calculating bootstrapped confidence intervals m ost
frequently used by applied statisticians, the percentile m ethod, is ideal fo r
use w ith the H a ll-V a n de Castle co n te n t indicators because it is consistent
w ith the use o f percentages fo r determ ining p values and effect sizes. In
add ition , the other m ethods e ith e r revert to param etric assumptions or
require hundreds o f thousands o f resamples, w ith little gain (M ooney &
D uval, 1993). A lth o u g h the percentile m ethod is o f lim ite d usefulness w ith
samples o f fewer than 30 and requires 1,000 or m ore resamples, these poten
tia l problems are n o t relevant fo r H a ll-V a n de Castle analyses because
they use large samples and rely on resam pling software available through
DreamResearch.net.
S tatistics textbooks usually suggest several hundred to 1,000 resamples
to determ ine the 95% confidence in te rva l. T o set the 95% confidence
in te rva l, the bootstrapping program available through DreamResearch.net

first generates a value fo r each o f 5,000 resamples o f the in d ica to r being
studied. N e x t, it counts dow n from the largest value in the d is trib u tio n to
the 125th largest value and then up from the sm allest value to the 125th
sm allest value. T h e 125 th largest value and the 125 th sm allest value are
the 95% confidence in te rv a l w ith 5,000 resamples. (In other words, 5% o f
5,000 equals 250, w ith h a lf o f the 250 at one end o f the d is trib u tio n , the
other h a lf at the other end.)
D eterm ining E ffe ct Sizes
C ritic s o f co n ve n tio n a l significance testing argue fo r an emphasis on

the m agnitude o f differences between tw o samples and greater use o f “ effect
size” statistics (R osenthal, Rosnow, & R ubin, 2000). T he H a ll-V a n de
Castle system responds to th is new emphasis by using C ohen’s h statistic,
w h ich is ideal fo r use w ith percentage data. C onceptually, the effect size
w hen com paring tw o percentages is sim ply the difference between them .
M oreover, th is difference is equal to p h i and lambda, the tw o statistics used
to determ ine effect sizes w ith c h i square. T h is result m ig h t be expected on
the basis o f th e earlier discussion o f the equivalence o f proportions and ch i
square w ith 2 x 2 tables (Ferguson, 1981; Reynolds, 1984). I t is also im p o rta n t
to note th a t the m agnitude o f the rela tio n sh ip between tw o percentages
can be understood in co rre la tio n a l term s: A s R osenthal and R u b in (1982)
show, the difference between tw o percentages is equal to Pearson’s r fo r
dichotom ous variables, so n o th in g is gained by using a co rre la tio n a l approach
instead o f percentages.
U n fo rtu n a te ly, an effect size using differences between tw o percentages
cannot be q u ite as sim ple co m p u ta tio n a lly as it is conceptually. A corre ctio n
has to be made fo r the fact th a t it is n o t possible to calculate the standard
d e via tio n o f a d is trib u tio n o f scores th a t has been transform ed in to percen
tiles. T o deal w ith th is problem , C ohen (1977) created the h statistic, w h ich
is based on an arcsine transform ation o f the percentages calculated fo r the
tw o samples. T h is corre ction can be made using tables found in C ohen
(1977, p. 183) and D o m h o ff (1996, p. 315). M oreover, the h sta tistic its e lf
is calculated by the D ream S A T spreadsheet on Dream Research.net, m aking
use o f the table unnecessary. G enerally speaking, h is a little m ore th a n
tw ice as large as the percentage difference between tw o samples w hen the
percentages fo r the tw o samples are between 20 and 80. I t is o n ly at the
extremes th a t h becomes increasingly large.
T he d e te rm in a tio n o f w hat is a sm all, m edium , or large effect size
varies from research area to research area and is, in large part, a judgm ent
th a t is based on experience. C ohen suggests th a t a sensible startin g p o in t
is to consider h = .20 a sm all effect size, h = .40 a m edium effect size, and
h = .80 a large effect size. However, he also advises researchers to “ avoid

the use o f these conventions” if they can substitute “ exact values provided
by theory o r experience” in the specific area in w h ich they w ork (C ohen,
1977, p. 184). T he experience w ith H a ll-V a n de Castle findings to date
has found th a t effect sizes up to .20 are sm all, effect sizes from .21 to .40
are m edium , and effect sizes above .40 are large. Effect sizes o f .50 or above
have been extrem ely rare except in a few case studies (D om hoff, 1996,
chapter 8; D om hoff, 1999b).
T o provide some perspective on the re lative m agnitude o f these effect
sizes, Table 3.6 presents the mean effect sizes th a t Rosnow and R osenthal
(1997) calculated fo r several areas o f psychological research, along w ith the
equivalent h values calculated fo r th is book. The table shows th a t the effect
sizes in H a ll-V a n de Castle studies are in the m iddle o f the range.
There is good reason to believe th a t h works ju st as w e ll w ith the
“ repeated measures” o f a long-term dream series as it does w ith independent
samples. T h is cla im derives from a sim u la tio n study using 10,000 resamples.
The sim ulation shows th a t C ohen’s form ula fo r determ ining the effect size
between tw o independent means, d, produced alm ost exactly the same results
w ith correlated measures as did a m ore com plex form ula th a t takes in to
account the size o f any corre latio n between tw o samples. Because proportions
are sim ply a special case o f the mean, th is result holds fo r C o hen’s h as
w e ll. A lth o u g h the authors concluded th a t the more com plex equation used
in th e ir sim ulation study is “ consistently slig h tly more accurate,” they also
noted th a t “ the differences are quite sm all and are triv ia l fo r the sample
size o f 50” (D unlap, C o rtin a , Vaslow, & Burke, 1996, p. 172). Because the
differences are sm all and the sample sizes are alm ost always greater than 50
in H a ll-V a n de Castle studies, it is feasible to use h fo r b o th independent
and repeated-measures studies o f dream content.
T he h statistic also makes it possible to create a clear “ graphic represen
ta tio n ” o f the k in d C ohen (1990) advocates because the effect sizes fo r any
TABLE 3.6
Effect Sizes in Selected Areas of Psychological Research
Mean effect size Mean effect size Equivalent effect size
in P earson’s r in C ohen’s d in C oh e n ’s h
Laboratory interview s .07 .14 .14
R eaction tim e .08 .17 .16
H a ll-V a n de C astle
studies .2 0 -4 0
Learning .26 .54 .52
Person perception .27 .55 .54
Inkblot tests .39 .84 .78
Everyday situations .40 .88 .80
Note. From People Studying People (p. 47), by Ft. L. Rosnow and R. Rosenthal, 1997, New York: W. H.
Freeman and Company. Copyright 1997 by W. H. Freeman and Company. Adapted with permission.
89
subset o f the H a ll-V a n de Castle indicators can be placed on a bar graph
th a t resembles an M M P I profile . T h is “ h p ro file ” im m ediately reveals the
com parative size o f the h differences between any tw o samples on the co n te n t
indicators th a t are being used. W h e n new samples are regularly compared
w ith the norm s, it is possible to determ ine q u ickly w hether any consistent
patterns exist fo r p a rticu la r types o f in dividuals o r groups. Patterns th a t
are based on com parisons o f h profiles are the closest app ro xim ation to a
m u ltiva ria te analysis th a t can be achieved w ith the H a ll-V a n de Castle
co n te n t indicators.
H ow ever, the h statistic cannot be used to determ ine effect sizes fo r
th e A /C and F/C indexes because h is based on a m athem atical transform a
tio n th a t is v a lid o n ly fo r distribu tions th a t vary between 0 and 1. T he
sim ple a rith m e tic difference between the tw o sample ratios is therefore used
as the basis fo r estim ating the effect size. Because the A /C and F/C ratios
are between .20 and .80 in m ost samples, the la ck o f a corre ction fo r extrem e
scores in th is m ethod o f determ ining effect sizes is n o t a problem . T he
effect-size findings fo r these tw o indexes are made com parable to h by
m u ltip ly in g by 2.36, a figure th a t was determ ined through an em pirical
analysis using m any subsamples o f H a ll-V a n de Castle indicators.
Figure 3.4 presents the h-profile fo r a com parison o f the m ale and
fem ale norm s; the p ro file is based on the m ethods described in th is chapter
fo r determ ining p values and effect sizes, using the fem ale norm s as the
“ baseline” group. T h e largest effect sizes concern the higher m ale/fem ale
and physical aggression percents, the higher percentage o f dreams w ith at
least one sexual in te ra ctio n , and the low er fa m ilia r settings percent in m en’s
dream reports. W ith regard to the guidelines fo r sm all, m edium , and large
effect sizes stated earlier, there is only one large difference: the h ig h e r m ale/
fem ale percent in the dreams o f men. For those w ho w ant to study gender
differences, it also can be said th a t there is a p a tte rn o f sm all differences;
b u t fo r purposes o f developing a neurocognitive m odel o f dreams, w h ich is
the goal o f th is book, the im p o rta n t p o in t is th a t the norm s provide a
baseline fo r com paring new samples o f m en or w om en and fo r studying
in d iv id u a l m en and women.
T he Im portance o f R eplications
A cco rd in g to C ohen (1990), there is no substitute fo r re p lic a tio n

studies in psychology no m atter how large the sample size or how great the
soph isticatio n o f the sta tistica l analysis. T h is a d m o n itio n seems doubly
im p o rta n t fo r dream research, in w h ich co n tra d icto ry results are frequent
because o f the m any sm all studies using d iffe re n t ra tin g scales. I t therefore
should be taken as a ru le th a t any result reported in the dream lite ra tu re
is considered te n ta tiv e u n til it has been replicated at least once. T h is rule

h
-.50 -.40 -.30 -.20 -.10 0 +.10 +.20 +.30 +.40 +.50
Male/Female %
Familiarity %
Friends %
Animal %
Aggression/Friendiiness %
Befriender %
Aggressor %
Aggression/Character Index
Fn'endliness/Character Index
Indoor Setting %
Familiar Setting %
Self-Negativity %
Bodily Misfortunes %
Negative Emotions %
At Least One Aggression

At Least One Friendliness
At Least One Sexuality
At Least One Misfortune
At Least One Success
At Least One Failure
*£><.05. * * p < .01,
Figure 3.4. T he h profile fo r the H a li-V a n de Castle m ale norm ative sam ple, using
the fem ale norm ative sam ple as a baseline. For instance, m en are 26 h points low er
than w om en in the percentage of hum an dream characters th a t are know n to them ,
w hereas they are 33 h points higher on the percentage of th e ir aggressions th a t are
physical in nature.
91
makes it possible to b u ild on solid findings in w o rkin g tow ard the u ltim a te
goal— th a t is, a good theory— rath er th a n spend tim e c ritiq u in g each riv a l
study or developing u n lik e ly explanations th a t span the range o f dispa
rate findings.
N EC ESSAR Y S A M P LE SIZES
C ohen’s (1990) tre n ch a n t critiq u e o f th e use o f statistics in psychologi

cal research also notes th a t sm all sample sizes are a m ajor problem th a t
leads to m isleading results. T h is problem looms even larger in the study o f
dream co n te n t fo r tw o reasons. First, the crucial issue fo r studies o f dream
co n te n t using no m in a l categories is the frequency w ith w h ich various catego
ries appear per dream report, n o t the to ta l num ber o f dream reports. If, fo r
exam ple, frie n d ly interactions o n ly occur in 42% o f w om en’s dreams and
37% o f m en’s dreams, as the norm s show to be the case, th e n a sample o f
30 dreams actually contains far fewer th a n 30 frie n d ly in teractio ns. Second,
the re la tiv e ly sm all differences th a t exist between in d ivid u a ls and groups
on m ost co n te n t indicators cann ot be detected w ith sm all samples w hen
co n ve n tio n a l tests o f significance are used.
T w o earlier studies dem onstrated th a t the average m in im u m sample
size necessary fo r studies using th e H a ll-V a n de Castle system is 100 to 125
dream reports if a ll the indicators are to be used. T h e first study m aking
th is p o in t is based on m any thousands o f subsamples o f 25, 50, 75, 100,
and 125 dream reports drawn fro m th e o rig in a l codings o f the 500 dreams
used to establish the m ale norm s. A de te rm in a tio n o f “ average departures
from the norm s” fo r each subsample size revealed th a t subsamples did n o t
regularly approxim ate the norm ative figure fo r m ost co n te n t categories u n til
at least 100 dream reports were included (D om hoff, 1996, pp. 6 5 -6 6 ). A t
the same tim e, 250 reports replicated the norm s perfectly, and samples w ith
125 reports came close. In a second study, using a 3 -m o n th dream jo u rn a l
kept by a n a tu ra l scientist, the results w ith the 178 dream reports w ith 50
o r m ore words were compared w ith those fro m subsamples o f various sizes.
T h is study showed th a t it to o k 100 dream reports to come w ith in 5 percentage
points o f the o ve ra ll findings fo r m ost indicators (D om hoff, 1996, p. 148).
Because p values are a fu n c tio n o f b o th the size o f the difference
between tw o groups and the sample size, it is possible to estim ate the
approxim ate num ber o f dream reports th a t are needed fo r o b ta in in g s ig n ifi
cant p values. For example, in testing fo r the significance o f the difference
between the m ale/fem ale percents in tw o samples, each sample has to have
100 dream reports if an h difference o f .20 is to be sig n ifica n t at the .05 level.
Each sample w ould have to have 175 dream reports to reach significance at
the .01 le vel. H ow ever, if the h difference fo r the m ale/fem ale percent is
92 THE SCIENTIFIC STUDY OF DREAMS 9

.40, as it is between the m en and wom en w ho contribu ted to the norm ative
samples, then on ly 25 dream reports w ould be needed in each sample.
T he results in the previous paragraph were obtained by com bining the
findings w ith the 1,000 dream reports in the male and female norm ative
samples to determ ine the mean num ber o f elements th a t appeared in each
dream report fo r each H a ll-V a n de Castle in d ica to r. T h is element rate makes
it possible to calculate approxim ately how m any dream reports are needed
fo r significance at the .05 or .01 levels fo r each in d ica to r, depending on
the size o f h. T he calcula tion involves the rate o f elements per dream,
the relevant z score (1.96 or 2.54, fo r significance levels o f .05 and .01,
respectively), and the h yp o th e tica l h value. Selected findings from this
analysis are presented in Tables 3.7 and 3.8. Table 3.7 shows the sample
sizes needed fo r significance at the .05 le vel w hen h is .20, .35, or .50; these
values encompass the range fo r h in m ost dream studies. Table 3.8 presents
the h difference th a t is necessary to obtain significance at the .05 le vel w ith
samples o f 30, 75, and 125 dream reports.
I f these results are taken as seriously as they should be, th e n the
shortcom ings o f the dream research lite ra tu re should be readily apparent.
T A B L E 3 .7
E s tim a te d M in im u m N u m b e r o f D re a m R e p o rts N e e d e d
in E a ch C o m p a ris o n G ro u p to Find a S ta tis tic a lly S ig n ific a n t D iffe re n c e
at th e .05 L eve l
C ritical n needed
fo r these hypothetical
Estim ated rate
h differences
o f elem ents
per dream. h = ± .2 0 ’ * h = ± .35 h = ±.50
Characters
M ale/fem ale percent 1.927 100 33 16
Fam iliarity percent 2.471 78 26 13
Friends percent 2.471 78 26 13
Fam ily percent 2.471 78 26 13
A ggression/friendliness percent 1.076 179 59 29
A ggressor percent 0.484 397 130 64
Physical aggressiofi percent 0.739 260 85 42
Settings
Indoor setting percent------------------ 1.177 164 54 27
Fam iliar setting percent 0.626 307 101 50
Note. T he fo rm u la used to calculate th e se "critical n” va lu e s is as follow s:
Z c is the critical Z -score fo r sig n ifica n ce a t the desired p level (in th is table, Zc= 1.96), h is th e hypothetical
h d ifference betw een the tw o gro u p s being com pared, and r is th e rate at w h ich th e coding e le m e n ts in
que stion ty p ic a lly a p p e a r in a single dream .
T A B L E 3.8
E s tim a te d h D iffe re n c e N e e d e d to F ind a S ta tis tic a lly S ig n ific a n t D iffe re n c e
a t th e .05 L e ve l, G iv e n T w o D re a m S e rie s o f E q u a l Len g th
C ritical h needed
E stim ated rate
fo r these sam ple sizes
n = 75
CO
o
per dream
c
n = 125
II
Characters
M ale/fem ale percent 1.927 .36 .23 .18
Fam iliarity percent 2.471 .32 .20 .16
Friends percent 2.471 .32 .20 .16
Fam ily percent 2.471 .32 .20 .16
A ggression /frie ndliness percent 1.076 .49 .31 .24
A ggressor percent 0.484 .73 .46 .36
Physical aggression percent 0.739 .59 .37 .29
Settings
Indoor setting percent 1.177 .47 .30 .23
Fam iliar setting percent 0.626 .64 .40 .31
Note. T he fo rm u la used to ca lc u la te th e se "c ritica l h" va lu e s is as follow s:
Zc is th e critical Z -sco re fo r s ig n ific a n c e a t th e desire d p level (in th is table, Zc= 1.96), n is the n u m b e r of
dre a m reports in th e tw o s a m p le s b eing c o m pared (it is a ssu m e d th a t the. sa m p le s a re o f e qual size), and
r i s th e rate a t w h ic h th e coding e le m e n ts in que stion typ ica lly a p p e a r in a sin g le dream .
First, sim ila ritie s from group to group and person to person may have been
overemphasized in some studies because o f sm all sample sizes. Second, m any
interesting hypotheses probably have been rejected prem aturely. Thus, large
sample sizes are needed to construct a sound theory.
C O N C L U S IO N
T he H a ll-V a n de Castle system o f co n te n t analysis is com prehensive,

reliable, and easily adapted to m odem sta tistica l procedures. I t has yielded
sound findings on how dream co n te n t relates to age, culture, psychopathol
ogy, and in d iv id u a l differences w hen applied to samples o f adequate size,
and it has show n th a t dream co n te n t is generally consistent over tim e
and continuous w ith e ith e r past or present w aking em otional concerns.
Furtherm ore, as chapter 4 shows, it now can be supported and supplem ented
w ith th e aid o f fast searches th a t use a W eb site on the In te rn e t.

A NEW RESOURCE FOR
CONTENT ANALYSIS
T he biggest drawback w ith n o m in al categories such as those used in

the H a ll-V a n de Castle system is th a t they are labor intensive. Part o f this
problem has been solved by D ream SA T, w h ich provides in sta n t and accurate
data analyses once the codings are entered. S till m ore o f the problem can
be m itigated through the developm ent o f Internet-based resources th a t
are accessible to a ll dream researchers. Such o n lin e resources have several
advantages th a t stem from centralization: They can be used from any lo cation
w ith o u t the need fo r extra software, o n ly one database needs to be
m aintained, and any im provem ents to the program o r additions to the
database benefit everyone.
T h is chapter discusses one such resource, http://w w w .D ream B ank.net,
w h ich was created by Adam Schneider in 1999. T h is site contains both a
large c o lle c tio n o f dream reports and a search engine, w h ich is w ritte n in
the Perl program m ing language. T he site can be used in at least three ways.
F irst, it can be used to id e n tify subsamples in long-term dream series. Second,
it can be used to conduct consistency studies fo r specific characters and
elements in a dream series. T h ird , it can be used to explore possible new
categories fo r doing content analysis w ith the search engine.
T he search engine locates in d iv id u a l words, long strings o f words, or
phrases by means o f regular expressions, w h ich are codes used fo r pattern
m atching in com puter program m ing (see Schneider &. D om hoff, 1999, fo r
a detailed discussion o f regular expressions). O nce the pattern is located,
the search engine reports the id e n tific a tio n num ber fo r each dream th a t
contains the requested query, along w ith the percentage o f dreams containing
the words. C ontingencies between patterns also can be calculated, such as
the degree o f relatio nship between the words “ house” and “fa th e r.”
T he dream reports retrieved by the search engine can be viewed in
fu ll on the screen or in an abbreviated form th a t displays o n ly the sentences
w ith the requested words in them . T he requested words are h ig h lig h te d and
in boldface w hen the dreams are brought up on the screen fo r view ing.
W h e n scrolling through the dreams on the screen, those th a t are n o t relevant
95
can be elim ina ted before the dreams are p rin te d o r analyzed on the screen.
I t is also possible to draw random samples from Dream Bank search results.
T h e m inim u m and m axim um num ber o f words per dream in the random
sample can be specified along w ith the desired sample size.
T he site contains m ore th a n 11,000 dream reports from groups and
in d ivid u a ls o f a ll ages. T h e reports can be used fo r a w ide range o f studies,
in c lu d in g those re la tin g to figurative th o u g h t in dreams. They can be draw n
upon to create new sets o f dreams on specific topics, such as dreams th a t
co n ta in weddings, bridges, o r murders. Researchers who w ant to use the
search engine to study dream reports they have collected them selves b u t
do n o t w a nt those reports to be p a rt o f the Dream Bank site may arrange
fo r a co n fid e n tia l site. A lth o u g h every e ffo rt has been made to ensure th a t
the dream reports on D ream Bank.net are accurate and authen tic, some o f
the dreams from ch ild re n and teenagers may be the product o f poe tic license.
T he best defense against th is p o ssib ility is to use the largest sample size th a t
is feasible. Table 4-1 presents an overview o f the dream reports available
on Dream Bank.net.
F A C IL IT A T IN G A H A L L -V A N DE C A S T LE
C O N T E N T A N A L Y S IS
T h e coding o f large dream samples using the H a ll-V a n de Castle

categories can be aided in several ways th a t save m any hours o f reading
thro u g h dream reports. Sim ple w ord searches are m ost valuable in fa c ilita tin g
the coding o f social in teractio ns w ith specific dream characters. For exam ple,
in the 3,116 dreams o f the Barb Sanders series analyzed in chapter 5, one
o f her brothers appears 97 tim es and one o f her daughters appears 165 tim es.
H e r aggressive and frie n d ly social in teractio ns w ith them were coded at the
rate o f one o r tw o per m inute by entering th e ir names in to the search
program and scro llin g through the dream reports onlin e.
W o rd strings are m ost useful w hen they are based on phrases or
expressions th a t are found to be frequent in an in d iv id u a l dream series.
For exam ple, m ost dreamers use on ly a few phrases fo r sexual in teractio ns,
such as “ made love” o r “ had sex.” These phrases th e n can be com bined
w ith a standard string o f sexual in te ra ctio n terms created by using “pipes,”
w h ic h are the means fo r sig nifyin g “ o r” in the language o f regular expres
sions. Thus, a string lik e “ kiss|hug|intercourse|m ade_love|had_sex” m ig h t
capture a representative sample o f sex dreams in a given dream series.
W o rd strings also can be useful fo r creating tw o subsamples th a t are based
o n d iffe re n t defin ing elements, w h ich can be compared w ith each other.
For exam ple, the frequency o f s triv in g and the percentage o f negative

T A B L E 4.1
T h e D re a m S e rie s A va ila b le on h ttp ://w w w .D re a m B a n k .n e t
G ender Y ears N
Alta: a detailed dream er female 1985-1997 302
Arlie: a m iddle-aged woman female 1992-19 98 212
Barb Sanders female 1960-1997 3,116
Barb Sanders: baseline female 1960-1999 250
Bay Area girls female 1996-19 97 388
Betty: a college student female 1993 28
Blind dreamers male and female m id-1990s 381
Chuck: a physical scientist male 1991-1993 75
College men and women,
1999 male and fem ale 1999 320
College women, late 1940s female 1946-19 50 667
David: teenage dreams male 1990-1999 156
Dorothea: 53 years of dreams female 1912-1965 900
Edna: a blind woman female 1948-1949 19
Emma’s husband male 1940-1998 72
Emma: 48 years of dreams female 1949-19 97 1,221
Female norms female 1940s-1950s 491
Jeff: a lucid dream er male 2000 87
Joan: a lesbian female mid-1980s 42
Male norms male 1940s-1950s 491
Mark: a young boy male 1997-1999 23
Melissa: a young girl female 1998-2000 67
Miami home/lab male 1963-1965 445
The Natural Scientist male 1939 218
Prospero: a blind man male 1970 202
Prudence: a literary woman female 1927-1998 209
Samantha: in her 20s female 1992-1999 63
Santa Cruz m ultiethnic male and female 1998 35
Seventh grade girls female 1996 69
Swiss children male and female 1989-1995 299
Teenage girls (W est Coast) female mid-1990s 89
Teenagers (Midwest) male and female 1998 194
Topwater: a violent man male 1965-1980 301
Univ. California Santa Cruz
women fem ale 1996 81
W edding dream s female 19 40s-195 0s and 1990s 65
em otions could be compared fo r dream reports w ith music- and sports-

related term s in them .
There are also some dream elements th a t m ost people describe w ith
com m on words or phrases. A n y dream report th a t describes the experience
o f flyin g under the dreamer’s own power is lik e ly to co n ta in the w ord “ fly ,”
“flo a t,” or “ a ir,” although those words also w ill appear in some reports th a t
are n o t about the to p ic under discussion. S im ila rly, m ost metamorphoses
o r abrupt scene changes are lik e ly to be captured by “ turned in to ,” “ changed
in to ,” “ become,” “ transform ,” and “ suddenly.” There w ill be m any tim es
A NEW RESOURCE FOR CONTENT ANALYSIS 97

w hen these words do n o t im p ly a m etam orphosis or scene change, b u t those
instances can be elim ina ted by reading through the dream reports before
doing the analysis.
D ream Bank also makes it possible in some cases to save tim e by entering
large strings o f words th a t are lik e ly to locate m ost o f the elem ents th a t
m ig h t be coded fo r a p a rticu la r H a ll-V a n de Castle category. Such strings
seem p ra ctica l fo r codings fo r food and eating references, em otions, and
anim als. H ow ever, even the longest o f strings is lik e ly to miss some dream
reports th a t co n ta in instances o f the coding category, so there is a trade-
o ff between efficiency and accuracy th a t m ust be taken in to account. T he
loss o f accuracy can be gauged in some cases by reading thro u g h random
samples o f the dream reports to determ ine w hat percentage o f the codeable
elem ents were missed. M oreover, it is s till necessary to lo o k at each instance
brought to the screen in order to e lim in a te those th a t do n o t meet the
coding rules fo r one reason o r another. T h is la tte r p o in t can be dem onstrated
here w ith a b rie f analysis o f the anim als th a t appear in the norm ative male
and fem ale dream samples.
For th is dem onstration, the dream reports coded by H a ll and V a n de
Castle as co n ta in in g at least one anim al were located by entering the code for
anim als— A N I— in to the SearchCodings program available thro u g h h ttp ://
Dream Research.net fo r use w ith already coded sets o f dreams. O nce the
actual anim als in the dreams were w ritte n dow n, th is string o f anim al terms
was th e n entered in to the query box fo r a search o f a ll the reports in the
tw o norm ative samples. The search produced 8 m ore male reports and 10
m ore fem ale reports th a n were coded fo r an anim al character. T h is overshot
stems from instances such as “soldiers in a /oxhole,” a discussion o f “ sm all
words lik e ‘cat’ ,” and the presence o f a “ goose-necked lam p.” M oreover, the
search retrie ved figurative expressions, such as “fly in g lik e a bird,” “ m aking
pigs o f them selves,” and “ acting lik e maddened animals.” O nce those cases
were rem oved, the findings were sim ila r to the norm ative H a ll-V a n de
Castle figures fo r the percentage o f dream reports th a t have at least one
anim al. For futu re studies o f anim als in dreams, long strings o f anim al words
could be used as a starting p o in t, b u t they w ould have to be augmented by
a reading o f the dream series under study.
U S IN G D R E A M B A N K .N E T FO R C O N T E N T A N A L Y S IS
T he search engine on D ream Bank.net makes it possible to carry out

a ll fo u r steps o f a co n te n t analysis: (a) create carefully defined categories
th a t have h ig h re lia b ility , (b) determ ine frequencies fo r each category, (c)
transform th e raw frequencies in to percentages o r rates, and (d ) compare
the findings w ith norm s or a c o n tro l group. T h is approach is com pletely

independent o f the H a ll-V a n de Castle coding system. T he phrases or
strings o f words entered in to the search program in effect define the new
categories. Because the same results are guaranteed each tim e the sample
is searched, the categories have perfect re lia b ility . The search engine also
does the second step o f a content analysis, w h ich is to provide frequency
counts. T he program generates some o f the necessary percentages, but th is
step has to be supplem ented in m ost studies by separate analyses o f the k in d
to be presented next.
T he fo u rth step o f a con te n t analysis, a com parison w ith a norm ative
or c o n tro l group, can be carried o u t by draw ing on other sets or series in
the Dream Bank. For m any studies, the ideal com parative sets may be the
dream reports th a t H a ll and V a n de Castle used in creating the male and
female norm s discussed in chapter 3. U sing these dream reports as a baseline
gives a study some com parability w ith H a ll-V a n de Castle findings. H ow ever,
in some cases it m ig h t make more sense to use another series as the com pari
son group. For example, one or tw o dream series from adult w om en m ig h t
make good com parison groups fo r studying a new dream series fro m a w om an
o f about the same age. T he com parison samples can be searched fo r the
phrase o r word string at the same tim e the m ain sample is being searched.
These points can be dem onstrated w ith a sim ple study o f pet anim als
in dreams. The categories first determ ine the frequency w ith w h ich people
m ention the tw o m ost com m on house pets: cats and dogs. These findings
are th e n used to create percentage-based indicators th a t distinguish between
“ cat people” and “ dog people.” T o begin, the search program is p u t on the
“ O R ” mode. N e x t, tw o strings o f words are entered th a t include regular
expressions o f various kinds to ensure a focused and q u a lity search:
A(cat||kitten|kitty|kittie|feline)s?AA(dog|doggy|doggie|puppy|puppies|canine)s?A.
T he n the dreams o f tw o cat lovers, “ A lta ” and Barb Sanders, are

selected fo r searching, along w ith those o f a young g irl, Melissa, who loves
a ll anim als, and an older woman, A rlie , w ho has little or no interest in
anim als. In add ition , the dreams used to create the female and m ale norm s
for the H a ll-V a n de Castle system are selected to provide com parison groups.
T he male norm ative dream sample is included to see w hether a gender
difference exists in the preference fo r cats and dogs. T he search produces
the results th a t are presented in Table 4.2.
A s can be seen in the table, the results reveal the num ber and percent
age o f dreams th a t co n ta in a m e n tio n o f at least one cat, at least one dog,
and at least one dog o r cat. T he th ird colum n from the le ft, w h ich contains
the summary in fo rm a tio n fo r the presence o f a cat or dog, provides an
in d ic a tio n o f a general interest in house pets. It can be te n ta tiv e ly called the
“pet percent.” A s expected, A rlie , w ith o n ly 1.9% o f her dreams co n ta in in g a
cat o r a dog, is below the fem ale norm o f 3.9% ; the other three dreamers

T A B L E 4.2
F in d in g s on P e ts W ith S e le c te d D re a m S e rie s on D re a m B a n k .n e t
Cats Dogs “ Pet p e rc e n t’ “C at percent”
Fem ale norm s ( n = 491) 9(1.8% ) 11(2.2% ) 19(3.9% ) 45%

M ale norm s (n = 491) 2(0.4% ) 11 (2.2%) 13(2.6% ) 15%
A lta (n = 422) 55(13.0% ) 21(5.0% ) 67(15.9% ) 72%
Arlie (n = 212) 1(0.5% ) 3(1.4% ) 4(1.9% ) 25%
Barb S anders ( n = 250) 12(4.8% ) 8(3.2% ) 18(7.2% ) 60%
M elissa (n = 67) 9(13.4% ) 11(16.4% ) 18(26.9% ) 45%
are above the fem ale norm , suggesting th a t they are interested in p e t anim als.
T here is a sm all b u t triv ia l difference between the m ale and fem ale norm ative
dreams on the pet percent.
A n in d ic a tio n o f the degree to w h ich a person dreams about cats as
opposed to dogs is obtained by d iv id in g the to ta l num ber o f dreams w ith
at least one cat by the to ta l num ber o f dreams w ith at least one cat or at
least one dog. R ecalling the assum ption th a t frequency reveals the in te n sity
o f in terest or concern, th is “ cat percent” reveals th a t A lta and Barb Sanders
are m ore concerned w ith cats th a n dogs, whereas young M elissa seems to
be interested in b o th cats and dogs. T he norm ative sample fo r wom en
suggests th a t A lta and Barb Sanders are w e ll above the average wom an on
cat percent. T he com parison o f the female and male norm ative dream
samples shows a large difference between w om en (45% ) and m en (15% ).
T he results o f th is sim ple study, in the co n te xt o f the c o n tin u ity p rin c ip le
discussed in chapter 1, lead to a substantive p re d ictio n : T he pet percent
and the cat percent w ill be continuous w ith w aking interests and concerns
in fu tu re studies using these scales.
M ore generally, the tw o scales show the possibilities fo r creating scales
fo r parents, fa m ily members, and friends. They w ould by no means be perfect,
b u t it is easy enough to figure out ju st how accurate they are by m aking
com parisons w ith hand tallies. T h e ir ju s tific a tio n is th a t they make possible
overview studies o f lo ng-te rm dream series th a t w ould n o t otherw ise be
feasible. H ow ever, it rem ains essential to ta ilo r each scale on th e basis o f
a careful exam ination o f the dream series to w h ich it is to be applied.
T he developm ent o f useful scales th a t are independent o f the H a ll-V a n
de C astle system also can be dem onstrated thro ugh a “ sensory references”
coding scale th a t was created to study 372 dream reports from 15 b lin d m en
and w om en (H u ro vitz et al., 1999). T h is exam ple also provides an e ntry
p o in t in to the issue o f id e n tify in g possible instances o f figurative though t
in dreams, because m any sensory terms are used m etaphorically (e.g., “ I see
w hat you m ean” ; “ the taste o f v ic to ry is sweet” ; and “ th a t deal smells fishy
to me” ). In this study, the researchers com pared the dreams o f people who

were congenitally or adventitiously b lin d and exam ined differences in the
percentage o f visual, auditory, olfactory, gustatory, and ta c tile references.
A ll forms o f the words “ see,” “ saw,” “ w atch,” “ lo o k,” and “ n o tice ” were used
to provide a starting p o in t fo r coding visual imagery. A ll forms o f “ hear”
and “ liste n ” were used to locate possible auditory imagery. A ll forms o f
“ taste,” “sm ell,” “ arom a,” “ scent,” “ feel,” “ fe lt,” and “ touch” were used as a
startin g p o in t fo r the other three senses. T h is lis t o f words is based on a
careful reading o f the dream reports fo r a ll sensory terms.
A fte r the dream reports w ith possible sensory references were located
by the search program, tw o coders studied each boldface word and its co n te xt
to determ ine w h ich sensory m entions seemed to be lite ra l and w h ich m ig h t
be figurative. T he d is tin c tio n between the lite ra l and the figurative use o f
term s is im p o rta n t in m ost studies, b u t it is especially crucial in a study
lo o kin g fo r visual im agery in the dreams o f b lin d participants. T he sensory
references were d ivid e d in to three categories: visual, auditory, and taste/
sm ell/touch, and the percentage o f each type was determ ined. The findings
were analyzed in terms o f a “ taste/sm ell/touch percent” because earlier studies
showed it to be the distinguishing sensory feature in the dreams o f b lin d
people. A s expected, the taste/sm ell/touch percent was h ig h in the dream
reports o f the co ng enitally b lin d participants. O n the other hand, fo r tw o
wom en who lost th e ir sight after age 8, o n ly 1 o f 36 sensory references was
in the taste/sm ell/touch category. Five o f the w om en’s sensory references were
auditory, and the rest (86% ) were visual. The findings w ith the adventitiously
b lin d participants were sim ila r to those fo r sighted adults in large-scale
studies (Snyder, 1970; Zadra, N ielsen, & D onderi, 1998).
T he seeming visual exceptions in the dreams o f tw o congenitally
b lin d people turned out to be m etaphoric in nature; they used a com m on
m etaphor in w h ich “ know in g” or “ experiencing” is “ seeing” . (M a tlo ck,
1988; M a tlo ck & Sweetser, 1989; Sweetser, 1990). For example, a 52-
year-old woman reported a dream in w h ich she and her husband (also
b lin d since b irth ) visite d w ith Thom as Jefferson h im se lf on a trip to
M o n tic e llo . She firs t says in the dream report th a t Jefferson “ was glad to
m eet us and he d id n ’t care if we couldn’t see,” a statem ent th a t confirm s
her la ck o f vision. However, she th e n reports th a t Jefferson to o k them to
“ see” the plants in his garden, a statem ent th a t appears to be an instance
o f “ experiencing is seeing.”
Scales aside, Dream Bank.net provides a qu ick and pow erful way to
study consistency in dreams. Such studies are m ost clear-cut fo r the consis
tency w ith w h ich a specific character appears, b u t they can be done as w e ll
fo r objects such as cars or a ctivitie s such as ru n n in g . T he frequencies th a t
are found can be turned in to percentages in one o f tw o ways: by d iv id in g
the frequency fo r a given year by the num ber o f dreams fo r th a t year, or
by using the num ber id e n tifica tio n s fo r each dream to determ ine frequency

per 50 or 100 dream reports. A example o f such consistency studies is
provided in the fin a l section o f th is chapter.
S T U D Y IN G F IG U R A T IV E T H O U G H T IN D R E A M S
T he D ream Bank search engine, especially w hen used in c o n ju n c tio n

w ith the H a ll-V a n de Castle categories, provides an entree in to the d iffic u lt
problem o f studying figurative th o u g h t in dreams. T w o b rie f examples, w h ich
concern dreams about weddings and dreams about bridges, are provided
here and are o n ly m eant to suggest possibilities.
W edding dreams are o f p o te n tia l interest fo r tw o reasons. W om en
have m ore o f them th a n m en, and wom en experience m ore mishaps and
mistakes in them th a n m en do. In terms o f the H a ll-V a n de C astle coding
system, previous studies suggest th a t wom en’s w edding dreams have an
unusually h ig h num ber o f m isfortunes— the w rong groom , the w rong church,
or the w rong dress, fo r exam ple (D om hoff, 1996). These m isfortunes may
be a useful startin g p o in t fo r the question o f figurative th o u g h t in dreams
because m any o f the phrases th a t are characterized as m isfortunes in the
H a ll-V a n de Castle system can be construed as m etaphoric (e.g., “ I was
lo st” or “ I was overw helm ed by a tid a l wave” ). T he p o ssib ility therefore
arises th a t w edding dreams m ig h t provide a co lle ctive m etaphoric portrayal
o f how A m e rica n w om en th in k about weddings.
T he firs t step in such a study w ould be to search a ll the adu lt dreams
on D ream B ank.net using a w ord string lik e w edding|m arried|m arriag'
e|groom |bride. T he search w ould provide a frequency fo r each in d iv id u a l
and group. A s a second step, random samples o f w edding dreams could be
created fo r b o th in d ivid u a ls and groups and th e n coded w ith the M isfortunes
scale. T h e n the findings fo r m en and wom en could be com pared w ith each
other and w ith the m ale and fem ale norms. I f the previous fin d in g o f
more m isfortunes in w om en’s w edding dreams is upheld, m ore detailed
and q u a lita tiv e studies o f the specific m isfortunes in those dreams could
be undertaken.
Dreams about bridges are o f p o te n tia l use fo r tw o reasons. First, they
are ce n tra l to some w aking m etaphors fo r personal tran sition s, such as “ cross
th a t bridge w hen you come to it , ” “ don’t bum your bridges,” and “ th a t’s
water under the bridge.” Second, there are few or no synonyms fo r the w ord
th a t designates these structures. C ircu m sta n tia l evidence th a t bridges may
serve as m etaphors fo r personal transitions in dreams can be found in a
detailed discussion o f “ bridgeness” in a dream report by Boss (1958), a
phenom enological dream th e o rist w ho professed to be adam antly opposed
to the idea o f sym bolism . I f bridges have a m etaphoric m eaning in dreams,
then perhaps they w ould be m ore lik e ly to be associated w ith m isfortunes
and negative em otions th a n they are in w aking life , where they are m atter-
of-fact, easily crossed, and seldom in danger o f breaking. So, after e lim in a tin g
the re la tive ly few instances o f the w ord “ bridge” th a t designate a card game
or dental w ork, the contingencies between bridges and m isfortunes and
between bridges and apprehension could be determ ined. I f the results o f an
earlier study using a sm all sample are any in d ica tio n , it is lik e ly th a t re la tio n
ships w ill be found (H a ll & N ordby, 1972, pp. 136-137).
U S IN G D R E A M B A N K .N E T T O S T U D Y T H E E M M A SERIES
M any o f the points made in th is b rie f overview can be dem onstrated

w ith several analyses o f the “ Emma” series. T he findings w ith th is series
also serve as an in tro d u c tio n to the extensive analysis o f the Barb Sanders
series in chapter 5.
Emma is a 77-year-old wom an w ho firs t w rote dow n her dreams w hen
she was in a Jungian analysis as a young woman. Several years la ter, when
m arried and the m other o f young ch ild re n , she began w ritin g dow n her
dreams again, w ith the in te n t o f studying them fo r sym bolism and archetypes
at some la te r p o in t. She eventually lost interest in analyzing the dreams
b u t continued the jo u rn a l out o f h a b it. T he jo u rn a l contains 1,221 dreams
in a ll, m ost o f them w ritte n down between ages 40 and 77. T here are no
dreams fo r 1981, 1988, and 1989, and o n ly 47 dreams fo r 1993 to 1997,
w hen the series ends. H er dreams came to the Dream Bank because she
asked a Jungian author fa m ilia r w ith co n te n t analysis if anyone m ig h t w ant
them . Emma w rote dow n her dreams fo r another 3 years after she first gave
her dream journals to the DreamBank.
Reading through the part o f the jo u rn a l th a t is re la tive ly continuous
over 37 years, the m ost strikin g im pression is the frequency w ith w h ic h her
husband and the m in iste r at her church appear. T h is im pression led to a
search o f the dreams fo r the num ber o f appearances by various friends and
relatives. I t confirm s th a t her husband and m inister are by fa r the m ost
frequent characters in the series. H er husband appears in 30% o f the 1,137
dreams during th is tim e period and her m inister in 23%. T aken together,
they appear in 48.3% o f those continuous dreams. I f frequency is an in d ica to r
o f “ in te n sity,” as a ll past studies indicate, th e n these tw o m en are her greatest
concern. Because Emma knew b o th o f them over the e n tire tim e period,
the question arises as to how consistently she dream t about them . T his
analysis shows th a t she was more lik e ly to dream about her m in iste r in each
o f the years from 1960 to 1971 and was increasingly lik e ly to dream about
her husband u n til 1990, after w h ich the tw o m en appear m ore equally in
the sm all num ber o f dreams th a t were recorded. Table 4.3 summarizes
the results.
T A B L E 4 .3
N u m b e r o f D re a m R e p o rts T h a t In clu d e E m m a ’s H u s b a n d o r M in iste r,
by T im e P e rio d
Y ears H usband M inister
1 9 6 0 -1 9 7 1 117 169
1 9 7 2 -1 9 8 9 201 70
1 9 9 0 -1 9 9 7 25 17
Total 343 256
A reading o f the dreams about the husband and the m in iste r suggests
th a t they do n o t o fte n appear in the same dream. T h is p o ssib ility can be
studied by using the contingency program th a t is part o f the search engine.
Because her husband appears in 30% o f the dreams and the m in iste r in
23% , the expected value fo r th e ir jo in t appearances is .069 (.30 X .23), or
78 dreams. In fact, they appear together o n ly 51 tim es. T h is difference has
a p value o f .02, suggesting th a t the tw o m en occupy som ewhat separate
spheres in the dream er’s m ind. A lth o u g h the effect size is n o t large, th is
fin d in g fits w ith the fa ct th a t Emma’s husband is n o t religious and does n o t
go to church w ith her.
T he question n a tu ra lly arises as to the nature o f Emma’s in te ra c tio n
w ith the tw o men, w h ic h can be explored by using the H a ll-V a n de Castle
categories fo r frie n d ly and aggressive interactio ns. T o answer th is question,
Thom as V a n Rompay firs t used the search engine to draw tw o random
samples o f 100 dreams th a t contained one o f the tw o dream characters.
T h e n the dreams were coded fo r frie n d ly and aggressive in teractio ns between
Emma and e ith e r o f the tw o men. T he contrast is strikin g : H e r in teractio ns
w ith her husband are aggressive, but her in teractio ns w ith her m in iste r are
frie n d ly. Furtherm ore, Emma usually in itia te s the aggressive in teractio ns
w ith her husband, w h ich m ostly consist o f angry thoughts, c ritic a l com m ents,
and ye llin g . She and the m in iste r b o th in itia te frie n d ly in teractio ns w hen
they in te ra ct. Some o f the specific findings from th is analysis are presented
in T able 4.4.
T A B L E 4 .4
E m m a ’s S o c ia l In te ra c tio n s W ith H e r H u s b a n d a nd M in is te r
H usband* (%) M inister* (%)
A t least one social interaction 26 52

P roportion of social interactions
th a t are friendly 11 93
P roportion of social interactions
th a t are aggressive 89 7
*n = 100 dre a m reports fo r each s am ple.

The findings w ith the tw o m en are a ll the m ore interesting in the
co n te xt o f the findings from a random sample o f 100 o f Emma’s dreams,
w h ich V an Rompay (2000) had coded earlier fo r another project. Com pared
w ith the fem ale norm s, the m ost atypical aspect o f her dream life is the
h ig h rate o f friendliness and the lo w rate o f aggression, w h ich makes the
pa tte rn w ith her husband even m ore striking.
Three inferences about Emma’s w aking thoughts can be draw n fro m
the analyses o f her dream series. First, Emma is intensely concerned about
b o th her husband and her m in iste r in w aking life , as shown by the large
num ber o f tim es they b o th appear in her dreams. Second, her interest in
the m inister, w h ich reached a h ig h p o in t w hen she was in her 30s and 40s,
has declined. T h ird , Emma has a greater num ber o f positive feelings about
her m inister th a n she does about her husband. Emma’s responses to questions
th a t were based on these inferences are discussed in the n e xt chapter, after
a m ethod fo r evaluating dreamer responses is introduced.
C O N C L U S IO N
T his chapter has dem onstrated some o f the possibilities fo r studying

dream reports by using the H a ll-V a n de Castle system and Dream Bank.net
in c o n ju n ctio n w ith each other. I t also has shown th a t D ream Bank.net can
be used to make analyses th a t are independent o f the H a ll-V a n de Castle
system. In add ition , the large num ber o f dreams th a t are available fo r
searching makes it possible fo r a w ide range o f researchers to study dreams
w ith o u t h aving to develop samples o f th e ir own. Researchers also could
w o rk together on in -d e p th studies o f a single series, such as the Barb Sanders
series discussed in chapter 5. A t the least, the dream reports on Dream -
B ank.net can serve as c o n tro l groups in studies by o ther researchers. T he
existence o f D ream Bank.net leads to the po ssib ility th a t fewer and fewer
studies in the dream lite ra tu re w ill have sm all sample sizes or use dream
reports th a t are shorter th a n 50 words in length.

5
NEW WAYS TO STUDY MEANING
IN DREAMS
T h e m ethods and discoveries discussed in chapters 1 th ro u g h 4 are a

s o lid fo u n d a tio n fo r fu tu re studies th a t a tte m p t to lin k dream c o n te n t to
b o th th e n e u ra l n e tw o rk fo r d ream ing and w a kin g c o g n itio n . Researchers
can n o w focus o n a ty p ic a l cases th a t m ake it possible to test specific h y p o th e
ses re la tin g to th e n e u ro c o g n itiv e m odel. T o do th is , researchers need good
dream recallers w h o are w illin g to supply ca n d id responses to inferences
d e rive d fro m b lin d analyses o f th e ir dream reports. T h e necessary dream
reports can be o b ta in e d in th e la b o ra to ry, th ro u g h hom e-based m o n ito rin g
using th e N ig h tc a p , o r fro m personal dream jo u rn a ls.
There are several reasons w hy in -de pth analyses o f in d iv id u a l dream
series, in co n ju n ctio n w ith neuropsychological o r neuroim aging studies, may
be the best research strategy fo r the developm ent o f the neurocognitive
m odel. First, as argued in chapter 2, experim ental studies o f dreams have
been o f lim ite d usefulness. Second, it is d iffic u lt to co lle ct a large representa
tiv e sample o f thorough dream reports from a group o f people o r to o b ta in
the am ount o f nondream in fo rm a tio n from them th a t is usually needed.
T h ird , the many findings on the relatio nship between dream co n te n t and
age, gender, culture, and psychopathology provide a co n te xt w ith in w h ich
specific hypotheses can be tested through detailed exam inations o f themes
and coding categories in a dream series. Fourth, a long-term dream series
can be studied expeditiously by m any d iffe re n t people from m any d iffe re n t
angles w ith the help o f Dream Bank.net.
As useful as in d iv id u a l dream series can be, most o f the early case
studies have lim ita tio n s o f one sort o r another. A lth o u g h they showed th a t
in fo rm a tio n can be extracted from a series o f dreams, the case studies did
n o t reveal w hether a ll aspects o f dreams are coherent and sensible. T hey
dem onstrated th a t many aspects o f dreams are continuous w ith w aking
personal concerns o f the past or present, b u t they d id n o t show w hether or
how a ll elements in a dream relate to w aking cogn ition . M ost o f a ll, they
d id n o t include a way to deal w ith disagreements between the inferences
o f the dream researcher and the replies o f the dreamer.
S T U D Y IN G A D R E A M SERIES
The purpose o f th is chapter is to present a new approach to studying

in d iv id u a l dream journals th a t m ig h t make it possible to understand dream
co n te n t in great d e ta il and to determ ine the degree to w h ich it is based in
figurative though t. T h is approach is dem onstrated w ith several d iffe re n t
analyses o f a series o f 3,116 dreams w ritte n dow n by an adu lt w om an know n
by the pseudonym “ Barb Sanders.” T h e study is, in many ways, the m ost
detailed e xam in ation o f a dream series th a t ever has been undertaken, but
it is n o t an exhaustive exam ination o f the series. There are m any m ore
analyses th a t w ould need to be conducted to provide a com plete p ictu re o f
the dreamer and her dreams. T o make such investigations possible, a ll the
dreams in the series are available to o th e r researchers on D ream Bank.net,
along w ith the codings fo r several H a ll-V a n de Castle categories in more
than 400 o f the dream reports. Future researchers also can apply th e ir ow n
coding systems to the dreams or use the search engine on D ream Bank.net
to study aspects o f the series th a t the present study does n o t cover. In
add ition , the results o f interview s w ith the dreamer and several o f her friends
are available on request.
In the several b lin d analyses o f dream series carried out by H a ll, replies
from the dreamers were the m ain source o f in fo rm a tio n concerning the
inferences made on the basis o f the dream reports (D om hoff, 1996, chapter
8). T h is approach makes it possible to focus on w hat seems to appear m ost
d ire ctly in dreams, nam ely, concerns and interests, by developing questions
th a t are based on inferences th a t arise from the analysis o f the dream reports
themselves. T h is approach makes it possible to bypass personality tests,
w h ich so far have n o t proven to be fru itfu l in the study o f dreams (D om hoff,
1996). T he usefulness o f th is m ethod is dem onstrated by the frequency
w ith w h ich H a ll’s inferences were corroborated, leading to the idea o f a
c o n tin u ity p rin cip le .
A sole reliance on the dreamers themselves fo r in fo rm a tio n has tw o
m ajor problem s, how ever. Sometimes dreamers do n o t have an answer fo r
a question, and sometimes they provide an answer th a t is n o t co n vin cin g
on the basis o f past studies. D isconfirm ing replies are reassuring, in the sense
th a t they increase confidence in the affirm ative replies, but they also generate
a fu rth e r question. H ow ce rta in is it th a t the dreamers’ negative answers
should be used to re je ct an inference th a t seems to be based o n a solid
generalization, nam ely, the c o n tin u ity principle?
T he problem can be dem onstrated w ith the case o f “ L u c ille ,” a wom an
in her 50s w ho began sending H a ll her dream reports after reading one o f
his books. T h is series o f several hundred dreams over a 10-year period
revealed strong consistency from year to year and an h p ro file close to the
norms on m ost indicators (Schneider & D om hoff, 1995). For the m ost part,
I the few deviations from the norm s are consistent w ith her w aking life . For
example, L u c ille has a h ig h percentage o f dreams w ith a t least one frie n d ly
in te ra ctio n (h = + .37); in her dreams, she shows kindness to h e r fam ily,
her supervisor, and little children, behavior th a t m irrors her w aking behavior.
T hen, too, she is above the norm in dreams w ith at least one m isfortune
(h = +.45) as a result o f her dreams about the in firm itie s o f b o th her husband
and w ork supervisor, w ho did in fa ct suffer from in firm itie s.
T he analysis o f the L u cille series also reveals th a t she dreams about
some people fa r m ore th a n others and has different patterns o f friendliness
and aggression w ith each o f them , as shown in Table 5.1. These patterns
parallel her feelings about them in w aking life , w ith one im p o rta n t exception
th a t is discussed shortly. L u c ille dreams most often about her husband,
w hom she saw every day u n til his death about 4 years after she started to
send dreams to H a ll. She has a h ig h rate o f b o th aggressive and frie n d ly
interactions w ith her husband, as shown by the A /C and F/C indexes
(described in chapter 3, this volum e), and the dreams have a greater num ber
o f aggressive th a n frie n d ly interactions, as shown by an aggression/friendli-
ness percent o f 58. T h is fin d in g fits w ith the couple’s in te ra c tio n pattern
o f sm all annoyances and rejections in w aking life .
A lth o u g h L u c ille sees a great deal o f b o th her sisters, w ho liv e together,
she dreams fa r m ore o f sister X (110 appearances) th a n sister Y (53 appear
ances) and has a far higher rate o f b o th aggressive and frie n d ly interactions
w ith sister X th a n sister Y. L u c ille reports greater em otion al in volve m en t
w ith Sister X th a n Sister Y, w h ich fits w ith the dream content. In lik e
fashion, L u c ille is close to her daughter, o f whom she dreams frequently,
and has m any m ore frie n d ly th a n aggressive interactions w ith her, w h ich —
according to her replies to H a ll’s in q u irie s— is also the case in w aking life .
She also was close to her m other, w hom she visite d d a ily u n til the m other
T A B L E 5.1
L u c ille ’s S o c ia l In te ra ctio n s W ith th e M ain C h a ra c te rs in H e r D re a m s
N um ber of
C haracter appearances A/C index F/C index A /F (%)
H usband 134 .66 .47 58
S iste r X 110 .41 .43 49
D aughter 103 .25 .58 30
S upervisor 83 .42 .61 41
Sister Y 53 .10 .07 59
M other 51 .04 .30 12
B rother 8 .00 .00 n/a
Father 2 .00 .00 n/a
Note. A /C index = a g g re ssio n s divided by c h a ra cte rs; F /C index = frien d lin e ss d ivid e d b y ch a ra cte rs;
A /F % = agg re ssio n s d ivided by aggressions plus friendliness.
NEW WAYS TO STUDY MEANING IN DREAMS 109

died, and th e ir in te ra ctio n s were far more positive th a n negative in b o th
dreams and w aking life . L u c ille dreams least o fte n o f her b ro th e r and her
fa th e r and does n o t have any frie n d ly or aggressive in teractio ns w ith them .
H e r b rother lives in a d ista n t c ity , and she rarely sees h im . H e r fa th e r died
w hen she was 9 years old.
T he dream patterns m atch extrem ely w e ll w ith her replies concerning
her relationships to these fa m ily members in w aking life , b u t such is n o t
the case w ith her w o rk supervisor. T his m an appears m ore o fte n th a n a ll
b u t three o ther dream characters. H e is in vo lve d in m ore frie n d ly th a n
aggressive in teractio ns w ith th e dreamer, w h ich is a d iffe re n t p a tte rn from
th a t in v o lv in g her husband. O n one or tw o occasions, the dream er and
her supervisor are hugging o r kissing. These contrasting findings w ith the
supervisor and husband are rem iniscent o f a sim ila r situ a tio n in the Emma
series discussed at the end o f chapter 4, in w h ich it was found th a t Emma
had positive in teractio ns w ith her m inister and negative in te ra ctio n s w ith
her husband. In L u c ille ’s case, H a ll inferred th a t she had ro m a n tic feelings
tow ard her supervisor, b u t she strongly rejected th a t idea.
S im ila rly , it now can be added to the in fo rm a tio n presented in chapter
4 th a t Emma does n o t agree th a t she has strong negative feelings tow ard
her husband. Emma agrees th a t her husband and m in iste r are the tw o m ost
im p o rta n t people in her life and th a t she has positive feelings about her
m inister, especially in the 1960s w hen she dreamed o f h im m ore freq uently.
She is re lu cta n t to say an yth in g c ritic a l about her re la tio n sh ip w ith her
husband. She reports th a t they had times o f tension between them b u t th a t
those tensions were tra n sito ry and are largely in the past.
B u t how can we be ce rta in th a t Emma’s dreams about her husband
and L u c ille ’s dreams about her supervisor do n o t fo llo w the c o n tin u ity
principle? I t is easy enough to say th a t they have chosen to report w h at
they know is n o t the tru th or to claim , as c lin ic a l theories th a t are based
on defense m echanisms can readily ju stify, th a t the dreamers may have
repressed th e ir feelings in those specific instances. H ow ever, th a t also means
th a t inferences cannot be falsified, leading to the ju stifie d c ritic is m th a t the
th e o rist is always rig h t no m atter w hat the dream er says.
T here are tw o ways to deal w ith th is dilem m a. T h e first is to seek o u t
letters and diaries from the dream er’s recent past. T he second, and better,
approach is to pose the same questions asked o f the dream er to people w ho
know the dream er w e ll. I f those people are in general agreement, they can
serve as judge and ju ry in the case o f disagreements between the dream er
and the researcher. I f they agree w ith the dreamer, the researcher is wrong.
I f they agree w ith the researcher, then the dreamer is u n w illin g to answer
tru th fu lly fo r some reason o r has a blind spot on the issue. T he phrase b lin d
spot has been carefully chosen to avoid any im p lic a tio n o f w hat process or
processes m ig h t be leading to an inaccurate com prehension o f one’s feelings.

I t is also a m etaphor th a t expresses the com m on understanding th a t everyone
has beliefs or opinions on some personal issues th a t are at variance w ith
w h at others judge to be the case.
T h is study o f Barb Sanders makes use o f interview s w ith fo u r o f her
close wom en friends, w ho proved to be knowledgeable about her personal
relationships, interests, and p o te n tia l b lin d spots. A s already noted, the
summaries o f those interview s are available upon request to researchers w ho
are doing th e ir ow n study o f the series. However, the present study is lim ite d
in its dem onstration o f the power o f th is approach: There were no m ajor
disagreements between the dreamer and the researcher or between the
dreamer and her friends. T h e dreamer rejected some o f the researcher’s
inferences, b u t the inferences in volve d issues about w h ich the friends had
no in fo rm a tio n ; in any event, it was su fficie n tly clear th a t the inferences
were wrong.
A P O R T R A IT OF B A R B SAN D ER S
Bom in the 1940s and raised in a sm all tow n, Barb Sanders is the
oldest o f four ch ild re n . She has a brother 2.7 years younger, a second b ro th e r
4.7 years younger, and a sister 6 years younger. B o th o f her parents earned
college degrees at a sm all, denom inational college and worked a ll th e ir lives
in education and social w ork. H e r parents also had a strong interest in
music, and a ll the c h ild re n sang and played m usical instrum ents. Sanders
was an average h ig h school student who m arried after 1 year o f college and
had three daughters in the space o f 4.5 years. H e r husband was a good
student w ho earned an M . A . and then w ent to w ork in a technical profession
fo r a natural resources corporation. Sanders earned a B .A . degree in her
m id-20s from a state college and le ft her husband at age 30, w hen her
daughters were 7, 4.5, and 2.5. H er daughters stayed w ith her ex-husband,
and she returned to her hom e state, where she earned an M .A . in a help in g
profession and w orked in a com m unity college setting fo r several years. She
had several boyfriends after her divorce and never rem arried. She became
in vo lve d in lo cal theater productions as an actress and director, and she
developed a strong in terest in dreams and participated in dream groups.
W hen the interview ees were asked about Sanders’s possible b lin d spots,
they had on ly a few reactions, b u t tw o o f the four said th a t she has a b lin d
spot when it comes to understanding her relationships w ith men:
First Friend: “ I th in k it ’s generally w ith male relationships, she doesn’t

see them for what they are. She doesn’t see them for what they are at
the tim e th a t she starts a relationship. A nd she, I th in k th a t she fantasizes
more in the beginning than what it is, and tries to make it w hat she

wants it to be. A nd G in n y (her other closest friend) and I share tha t
understanding about her.”
Second Friend: “W e ll, 1 th in k probably in relationships w ith men, she

has a certain naivete about her. 1 th in k o f her more as like a young
teenager in relationships w ith men in her development . . . when it
comes to her own personal life , I would say there’s a certain naivete,
this certain b lin d spot, yeah.”
Barb Sanders reported th a t she has always been fascinated w ith dreams
but d id n o t start a dream jo u rn a l u n til a few years after her divorce, at a
tim e w hen she was having disturbing dreams, in clu d in g some re la tin g to
the divorce and its afterm ath. She hoped to gain in sigh t about herself from
keeping the jo u rn a l and perhaps enhance her efforts at creative w ritin g .
Soon after she began recording her dreams, she also included in the jo u rn a l
several dreams she had w ritte n dow n in her late teens. She also w rote dow n
more th a n 100 dreams from m emory. T he early dreams th a t she w rote dow n
at the tim e she had them are included in the dream series th a t is analyzed
here, b u t n o t the dreams w ritte n dow n from memory. A lth o u g h the jo u rn a l
is continuous fro m the late 1970s to 1996, the num ber o f dreams w ritte n
dow n o r entered in to the com puter in each year varies. Figure 5.1 illustrates
the num ber o f dreams by year.
A s w ith m ost dream series, the dreams vary in le ngth, b u t 66% are
between 50 and 300 words. S ixteen percent have fewer th a n 50 words; a
Years
Figure 5.1. T h e distribution o f Barb S a n d e rs’s dream reports by years.

disproportionate num ber o f the short reports come from the 1960s and
early 1970s. Figure 5.2 presents th e com plete d is trib u tio n o f dream reports
by length.
The analysis o f the Barb Sanders series began w ith a H a ll-V a n de
Castle con te n t analysis o f a random sample o f 250 dream reports (the
“ baseline 250” ) ranging in length fro m 50 to 300 words. T his baseline sample
was coded for Characters, Social Inte ractio ns, M isfortunes, Successes and
Failures, and Em otions by Sarah D u nn, M elissa Bowen, and H e id i B lock,
who worked in different pairs o n d iffe re n t categories. Several re lia b ility
checks yielded the high percentages o f agreement shown in Table 5.2. A ll
differences o f o p in io n were resolved by discussion among the coders to
provide a u n ifo rm result fo r entry in to D ream SA T.
The first im portant fin d in g is a m ethodological one: The codings from
any 125 o f the 250 dreams in the random sample replicate the results alm ost
exactly, as determ ined by the approxim ate random ization program available
to researchers o n request through http://w w w .D ream R esearch.net. However,
there are many deviations from th e overall results w ith subsamples o f 100
dream reports, and the d ro p -o ff is large at 75 and 50, dem onstrating once
again how risky it is to accept or re je c t hypotheses on the basis o f inadequate
W ord C ount
Figure 5.2. The distribution o f Barb S a n d e rs ’s dream reports by length.

T A B L E 5.2
R e lia b ility F ig u re s fo r th e C o d in g s o f th e B a rb S a n d e rs B a se lin e S a m p le
o f 2 5 0 D re a m s
C ategory A greem ents D isagreem ents % of A g re e m e n t
C haracters 425 38 92
Aggression 188 17 92
Friendliness 106 20 84
S exuality 11 0 100
M isfortune/G ood fortune 109 15 88
Striving 45 13 78
Em otions 185 25 88
sample sizes. For exam ple, th e A /C index o f .33 is usually between .30 and
.35 w ith random ization samples co n ta in in g 125 dream reports, but it ranges
from .27 to .39 w ith samples o f 75 dreams.
T he first substantive analysis concerns the consistency in the dream
series (T able 5.3). A com parison o f the firs t 125 dreams in the baseline
sample w ith the second 125 found the dream er to be consistent w ith in 5
or 6 percentage points in a ll but tw o or three categories. H e r m ale/fem ale
T A B L E 5.3
C o n s is te n c y in th e B a rb S a n d e rs B a s e lin e 2 5 0 S a m p le W h e n th e T w o
H a lv e s A re C o m p a re d
First half S econd half h P

Characters
M ale/fem ale percent 58% 49% -.20 .013*
F am iliarity percent 38% 33% -.10 .135
Friends percent 20% 11% - .2 6 .000**
F am ily percent 15% 20% +.12 .076
Anim al percent 07% 06% - .0 5 .430
A g g ression /frie ndliness p e rce n t 49% 49% -.00 .984
B efriende r percent 49% 56% +.16 .251
A g g re sso r percent 53% 48% -.10 .480
P hysical aggression percent 29% 33% +.08 .484
A ggre ssio n /ch a ra cte r index .31 .35 +.09 .408
F riendliness/ch aracter index .29 .35 +.14 .060
S exuality/cha racter index .09 .10 +.02 .817
Dreams with at least one
A ggression 52% 56% +.08 .526
Friendliness 53% 54% +.02 .899
S exuality 22% 18% -.0 8 .527
*significant a t the .05 level

**.significant at the .01 level

percent fe ll from 58/42 to 49/51, placing her close to the female norm s o f
48/52. I t is also interesting th a t she dreamed less o f friends and more o f
fa m ily members in the second h a lf o f the series. The findings on consistency
support several previous studies dem onstrating consistency in dream co n te n t
over decades (D om hoff, 1996, chapter 7). T he result is now so w e ll estab
lished th a t deviations from it are o f interest in developing a better theory
or in understanding a specific person. T he fin d in g can be used as one litm us
test for determ ining the adequacy o f any dream theory.
C haracters and Social In te ra ctio n s
Characters and social interactions are usually the m ost psychologically

revealing aspect o f a detailed con te n t analysis, and the Sanders series is no
exception. T he random sample contains 884 hum an characters, 3.4 per
dream report. Because the female norm ative figure is 2.7, this fin d in g suggests
th a t Sanders is more intensely in vo lve d w ith o ther people in her dreams
th a n the typ ica l wom an is. T h is presum ption is supported w hen the 679
social interactions are analyzed in terms o f rates per character, th a t is, the
A /C , F/C, and S /C ratios. As shown in Table 5.4, the dreamer is 19 to 24
h points higher th a n the norm ative figures fo r these three ratios. T h e effect
sizes are sm all b u t sta tistica lly significant; w hat makes them n o tew orthy is
th a t a ll three indicators are elevated (Bow en & D unn, 1999).
A lth o u g h Barb Sanders has a h igher rate o f social in te ra c tio n th a n
the norm ative sample, she is typ ica l in her A /F percent (49) w h ich means
th a t she is equally in vo lve d in frie n d ly and aggressive social in teractio ns
w hen a ll dream characters are considered. However, she does d iffe r in th a t
she is far m ore lik e ly to be the aggressor (50 to 33), w h ich is 36 h points
above the norm , a m oderate effect size. M ost o f these aggressions are angry
thoughts tow ard a person, c ritic a l com ments, o r rejections. She is about
average on befriender percent (53).
T A B L E 5 .4
B a rb S a n d e rs ’s S o cia l In te ra ctio n P e rce n ts C o m p a re d W ith
th e F e m a le N o rm s
Barb S anders Fem ale
baseline 250 norms h P
A ggression/character index .33 .24 .21 .000**
Friendliness/character index .32 .22 .24 .000**
S exuality/character index .09 .01 .19 .000**
A ggression/friendiiness percent 49% 51% - .0 5 .400
A ggressor percent 50% 33% +.36 .000**
B efriender percent 53% 47% +.11 .255
T o provide an in d ic a tio n o f her patterns o f in te ra c tio n w ith key people
in the dream er’s life , R yan Harvey coded a ll o f the frie n d ly and aggressive
in teractio ns in the e n tire dream series w ith her parents, fa vo rite brother,
daughters, and tw o close w om en friends. T he findings, w h ich are summarized
in Table 5.5, correspond to how the dream er and the fo u r interview ees
describe her relationships w ith these people in w aking life . For exam ple,
her m other is the m ost im p o rta n t and d iffic u lt person in Barb Sanders’s life .
Sanders provided the fo llo w in g portrayal o f her m other:
M y m other is an angry, isolating person, and she also has good things
too, don’t get me wrong. But she and I have had a personality clash as
long as I can remember. I feel th a t she keeps herself so distant th a t I
d id n ’t feel I was getting nurturing m other love. I to ld one o f my women
friends th a t the love o f my m other is like carrying a barbed blue baby
blanket, you know , w ith barbs in it. I t ’s supposed to be soft and cuddly
and loving, but in fact, she was sharp and c ritic a l and negative and
physically distant.
H e r m other appears in 239 dream reports, or 7.7% o f the to ta l, w h ich

is m ore than any o th e r fa m ilia r character. T he A /C ra tio w ith her is .70,
w e ll above Sanders’s .33 average in the baseline 250 fo r a ll characters. T he
A /F percent between them is 72, w e ll above the dream er’s norm ative figure
o f 49 (h = .48). T h is A /F percent is consistent over the e n tire series, as
shown when the dreams are divided in to thirds.
O n the o th e r hand, Sanders stated th a t she has a m ore positive a ttitu d e
tow ard her father, a cla im th a t was corroborated by her friends. He appears
in 213 dream reports, second on ly to her m other. T he A /C ra tio w ith h im
is .36, n o t far above her norm ative figure w ith a ll characters, and the F/C
is .37, once again s lig h tly above her norm ative figure. T he A /F percent
fo r th e ir in teractio ns is 50. A lth o u g h n o th in g stands out about Sanders’s
T A B L E 5.5
B a rb S a n d e rs ’s S o c ia l In te ra c tio n s W ith S ig n ific a n t P e o p le in H e r Life,
C o m p a re d W ith th e B a s e lin e 2 5 0 S a m p le
A/C F/C A/F A ggressor B efriender

N index index (%) (%) (%)
B aseline 250 884 .33 .32 49 50 53
M other 239 .70 .27 72 46 48
Father 213 .36 .37 50 47 42
O ld e st daugh ter 81 .51 .65 44 73 77
M iddle daughter 165 .92 .52 64 79 70
Y o u n g e st daugh ter 83 .36 .81 31 63 61
Favorite brother 97 .23 .69 25 59 60
Friend G inny 96 .26 .89 23 52 53
Friend Lucy 59 .39 .63 38 78 78

relationship w ith h im , the re la tio n sh ip is dram atically d iffe re n t from her
relatio nship w ith her m other.
Sanders’s m iddle daughter is alm ost as problem atic fo r her as her
m other. T h is daughter was 4.5 years o ld at the tim e o f the divorce and was
the ch ild most upset by it. A t age 14 she ran away from her fa th e r’s hom e
and came to liv e w ith Sanders. She d id poorly in school, cannot h o ld a
job, and suffers fro m severe psychological problems. She had a daughter
w hen she was a teenager and soon le ft her to Barb Sanders to raise. She
s till returns to liv e w ith Barb Sanders from tim e to tim e. Sanders worries
about her constantly, and there is great tension between them .
T h is daughter appears in 165 dream reports. T he A /C ra tio is .92,
even higher than w ith the m other, and the F/C ra tio is also h ig h (.52, w e ll
beyond Sanders’s average for a ll characters). M ost o f the frie n d ly in teractio ns
in vo lve Sanders h e lp in g her daughter in one way or another. Taken together,
the tw o ratios show a high rate o f in te ra ctio n between them . T h e A /F
percent is 64, w ith Sanders in itia tin g 79% o f the aggressive interactions
and 70% o f the frie n d ly interactions, fa r above her averages fo r a ll characters.
These indicators provide an accurate summary o f how Sanders conceives
o f th e ir relatio nship.
By contrast, Sanders dreams o n ly h a lf as often o f her oldest and youngest
daughters, who adjusted to the d ivo rce better, w ent to school in th e ir fa th e r’s
hom e state, saw th e ir m other p rim a rily during summer vacations, and liv e
norm al adult lives. T he A /F percents, 44 w ith the older daughter and 31
w ith the younger daughter, show th a t she has more frie n d ly th a n aggressive
interactions w ith b o th o f them, w h ic h reflects her m ore positive rela tio n sh ip
w ith them . A s w ith the m iddle daughter, Sanders is m ore lik e ly to be the
in itia to r o f b o th aggressive and frie n d ly interactions.
T he dream reports also capture Sanders’s positive relationships w ith
the favorite people in her life. For example, Sanders has great a ffe ction fo r
the brother closest to her in age. H e appears in 97 dream reports, w h ich is
the same as the to ta l fo r her o th e r tw o siblings com bined. T he A /F percent
fo r th e ir in teractio ns is 25, alm ost the m irro r opposite o f her in te ra c tio n
pattern w ith her m other. Positive patterns also are apparent w ith tw o wom en
friends, but differences emerge on some indicators, w h ich reflect her d iffe re n t
pattern o f in te ra c tio n w ith each o f them . Sanders m et her closest frie n d o f
long-standing, G in n y , when she returned to college fo r her m aster’s degree.
A fte r college G in n y m arried, m oved to a c ity more th a n 100 m iles away,
and raised a fa m ily, b u t she and Sanders rem ained in close touch. H er
husband and c h ild re n are also frie n d ly w ith Sanders, who has visite d th e ir
hom e frequently over the years. G in n y appears in 96 dream reports and has
an A /F percent o f 23, the m ost p o sitive balance w ith any know n dream
character. T he com radely nature o f th e ir relatio nship is seen in the fa ct
th a t they are equally like ly to in itia te frie n d ly o r aggressive interactions.

Sanders m et another close frie n d , Lucy, w hen Lucy was a student at
the com m u nity college where Sanders worked. Lucy, w ho is several years
younger th a n Sanders, is outgoing and dram atic, and she and Sanders soon
ended up w o rking together in musicals and th e a trica l productions. She
appears in 59 dream reports. Sanders is the big sister in th is relatio nship;
she gives Lucy instructions, helps her, and becomes annoyed w hen Lucy is
late or resists d ire ctio n . T h is pa tte rn is reflected in the fa ct th a t Sanders
in itia te s 78% o f the m any frie n d ly interactions between them as w e ll as
78% o f the re la tive ly few aggressive interactions.
These analyses o f Sanders’s dreamed social in teractio ns w ith several
o f the sig nifica nt people in her life are on ly a starting p o in t in understanding
her w aking conceptions o f them . H ig h ly detailed studies w ith any o f the
characters w ould be lik e ly to yie ld ric h findings because o f the large num ber
o f dream reports. T o dem onstrate th is claim , som ewhat m ore detailed analy
ses are n e xt presented on tw o d iffe re n t subsets o f dream reports. O ne subset
concerns her ex-husband, the o ther a m an she was infatuated w ith fo r nearly
a year in the m iddle 1990s.
T h e E x-H usband Dream s
Barb Sanders first m et H ow ard, her future husband, w hen they were
in h ig h school, where they had lockers near each other. T he y sometimes
danced or flirte d , but he had a steady g irlfrie n d . In her senior year, Barb
fe ll in love w ith D arryl, the person she s till considers the true love o f her life .
H ow ever, th a t relatio nship d id n o t w o rk out fo r reasons th a t she explained in
the in te rv ie w 1:
Okay, so then D arryl went o ff to the Navy and that kind of ended it? W hat
happened was th a t we decided we would n o t go steady, but we were
s till going to get married, but we could explore w ith other people, tha t
k in d o f thing. It kin d o f broke my heart, and I found out in an indirect
way th a t he, when we were going steady, he was going out w ith some
other woman at the N avy base, and so I fe lt te rrib ly betrayed. So I
started going out w ith Howard, and then there was a th ird gentleman
by the name o f Pete, and a ll three o f them asked me to m arry them,
so it was quite a. . . . But by then you didn’t want to marry Darryl? I was
so angry and fe lt so betrayed th a t at th a t p o in t, I don’t know w hat it
is about this, but I, when I said, “T h a t’s it, we’re done. I ’m dating these
other guys, you know ,” it ’s like , “ O h no, please take me back, you know,
and I ’l l be, forgive me, etc., etc.” But by then I was done, I was out,
and very, very angry and was n o t w illin g to trust him again. So he was
‘ Sentences in ita lics in this and subsequent in te rv ie w segments in this chapter are th e questions
asked by th e author.
going to buy the engagement ring and we were going to get married,
but tha t’s it. I w ouldn’t ta lk to him , wouldn’t answer his letters back.
A t this p o in t, H ow ard was attending a m ajor un ive rsity far from th e ir

hom etow n and was no longer seeing his h ig h school g irlfrie n d . H e and
Sanders reconnected w hen he came hom e fo r the summer, and they were
m arried after an 18-m onth courtship. A lth o u g h it is risky to rely on memories
o f 35 years earlier, especially after a p a in fu l divorce, Sanders recalls th a t
she harbored some doubts about the relationship even then. Perhaps this
is n o t surprising in the co n te xt o f the h ig h ly m ixed feelings she harbored
about D a rryl and th e end o f the relatio nship w ith h im . She also remembers
H ow ard as a person who was insensitive to her need fo r tenderness and
expression o f feeling, as a person w ho ju st wanted sex. She fe lt th a t th e ir
sexual interactions sometimes fe lt m ore lik e rape th a n seduction.
H ow ard appears in 164 o f the 3,116 dream reports. He is surpassed
on ly by her m other, father, and m iddle daughter in the frequency o f his
appearances.2 These dreams have a fa irly regular structure (B lo ck, 1999).
T hey usually start w ith Sanders n o tic in g th a t H ow ard is back, causing her
considerable apprehension or annoyance. A t the outset he is ofte n seeking
re co n cilia tio n , although on occasion she is the one who is th in k in g about
the p ossibility o f getting back together. A s the dream unfolds, H ow ard
usually tries to in itia te a sexual in te ra ctio n , thro ugh a touch o r a kiss, but
Sanders is e ith e r hesita nt or repulsed. Sometimes she is tem pted, b u t th e n
changes her m ind. For example:
July 13, 1976: Howard wanted back w ith me. I put my arm around
h im but it was a terrible effort. T hen he started to kiss me and I pulled
away, feeling sickened and disgusted. “Don’t ever touch me again,” I
said. I woke up feeling good th a t I can leave it, and sick tha t I spent
10 years w ith him .
November 26, 1980: Howard, hovering around, w anting us to be
together. I have a tig h t smile on my face. I keep turning away from
him , but I ’m tempted.
March 4, 1981: I ’m in a house, getting ready to go home. Howard
is there. He wants to kiss me and go home w ith me. I want to be nice
to him but I don’t w ant h im back. I f I ’m nice to him , he’ll move rig h t
in. I feel trapped.
2T h e analysis o f the ex-husband dreams extends th ro u g h the m id dle o f 1999, 3 years after Barb
Sanders gave th e first 3,116 dreams to the Dream Bank, w h e n she p rovided 376 m ore dream reports.
A l l o th e r analyses in th is chapter (except fo r th e D erek dreams discussed in the n e x t se ctio n ) are
based o n the o rig in a l series o f dreams so as to avo id th e p o ssib ility th a t th e dream reports m ig h t be
in flu e n ced by th e fact th a t th e dreamer was n o w h e lp in g w ith a research p roject. A n e xce p tion was
made w ith th e ex-husband and D erek dreams to see w h e th er a process o f gradual change co u ld be
detected. T h is decision added 32 a d d itio na l H o w a rd dreams fo r a to ta l o f 196.

September 9, 1990: I am in bed w ith Howard again. I t ’s been so long
since we’ve made love and he wants it real bad. I agree. I lie down and
try to pretend I ’m asleep. I am turned away from him . He slowly and
gently comes close to me. He’s afraid I ’ll p u ll away or refuse h im and
so he’s being very hesitant and careful. He licks my cheek. I feel some
revulsion. He touches my nipple. I want to cringe. I try to tolerate it
but can’t, it is so repugnant to me. I sit up, crying and sobbing. He is
heartbroken and runs from the room.
May 8, 1991: Howard and I are in bed. He wants to make love. He
tries to kiss me. I am repulsed and sad. I feel empathy and sadness for
h im because he wants it/m e so badly. I feel so sad for me too. I want
sex, love, and relationship, and to say “no” to h im means saying no to
myself as w ell. B ut I can’t get over the repulsion I feel. He cries and
says “ You used to be so beautiful,” meaning when I was younger. I am
surprised. Maybe he is fin a lly able to express his emotions. B ut as he
continues to ta lk he clearly stays in a logical, d ig ita l conversation. I
am disappointed.
In 1996 and 1997, the H ow ard dreams seem to become som ewhat
m ore benign. T he y appear to co n ta in m ore re fle ctio n and regret as w e ll as
m ore discussion between the tw o o f them . In one or tw o dreams, she even
entertains the idea o f re co n c ilia tio n , an idea fo r w h ich there was no basis
in her or his w aking re a lity. T he y liv e d far apart, never saw each other,
and rarely com m unicated. In ad d itio n , H ow ard had been rem arried fo r m any
years. T h is im pression o f a change in th e tone o f the H ow ard dream reports
is borne o u t by the large decline in th e A /F percent w hen they are divide d
in to four ch rono logical segments; p ro p o rtio n a te ly more frie n d ly in te ra c tio n
occurs in the la te r years. T h is result is shown in Table 5.6.
Sanders’s reflections on her feelings about H ow ard in M arch 2000
pa ra lle l the m a in themes in the dreams as w e ll as the changes, as shown
in these excerpts from the in te rvie w w ith " her:
In the earlier years the dreams were ju st exact experiences o f my life
w ith him , so it was re al-life stuff. He really did beg, and he really did
try to get me back in the relationship. He really wanted the relationship
to stay. I was unable to do tha t. M y anger at his closed em otional stuff
was very strong in the dreams, it was very strong in my life.
T A B L E 5.6
B a rb S a n d e rs ’s D re a m s o f “ H o w a rd ” :
C h a n g e s in th e A g g re s s io n /F rie n d lin e s s P e rc e n t O v e r T im e
Barb S anders 1st 2nd 3rd 4th
baseline segm ent segm ent segm ent segm ent
A g gression /frie ndliness
percent 49 57 59 61 34

I t took a lo t o f years o f evolution, b u t in fact I watched myself go
from extrem ely painful anger, hardly able to m ention the word marriage
or Howard w ith o ut practically spitting, just being dram atic about the
whole thing. T h at was at the beginning, and the nightmares o f Howard
wanting me, or finding myself back in the relationship. O h man, they
were just, oh. A n d now I have Howard-is-back dreams and they’re okay.
I had a couple more even in the year 2000. H e’s s till there. But it is
changed. A forgiveness phase happened. Years and years and years,
anger, anger, anger, and slowly over th a t tim e learning more about
myself, how I operate, how he operates, and through the dreams them -
selves I was able to let go o f a lo t o f the anger and get more in touch
w ith the sadness.
Then, in the m idst o f th is apparent softening o f a ttitu d e tow ard H ow -

ard, he died unexpectedly o f a heart attack in A p ril 1997, w ith o u t any
h isto ry o f serious illness. A t th is p o in t the dreams about h im seem even
m ore reflective, and sometimes include the awareness th a t he is dead, even
though she is in te ra ctin g w ith h im in the dream:
May 5, 1999: I am lying on a bed, trying to sleep. Howard is seated

near me. W e talk. He is being helpful, trying to help me understand
something th a t would be good for me to know. Later, he is lying in the
bed w ith me. I “wake up” in the dream and realize th a t Howard is dead.
He can’t be there in my bed. I turn and look. He is not there. Then
he talks to me. So now I ’m wondering if I really looked or if I dreamed
I looked. I am confused if he’s really there or if I am dreaming it.
Despite- th is tren d tow ard a more positive resolution o f her feelings

about H ow ard and the divorce, the negative dreams s till appear on occasion.
O n N ovem ber 13,1997, fo r example, about 6 m onths after his death, Sanders
has a dream th a t is sim ila r to the five fro m 1976 to 1991 excerpted earlier:
I am in bed w ith Howard and he feels sexual desire for me. He’s looking
at me; he groans and says you’re so beautiful. I move away from h im
and say, “ Please don’t.” He looks sad and says “Please.” A nd I say I ’m
sorry. I p u ll away from h im so we’re n o t touching and I say to him ,
sometimes I th in k about just le ttin g you do it, to give you relief, but I
just can’t, after a ll those 10 years o f marriage and I begin to cry. He
gets out o f bed and comes around and tries awkwardly to help me feel
better. I feel very sad.
Sixteen m onths later, on February 13, 1999, Sanders has a dream in

w h ich H ow ard suddenly appears and asks fo r sex. She th in ks about it, b u t
th e n “ I p u ll back and say ‘n o .’ ” He th e n “ gets angry and grabs m y arm and
forces me over the w a ll o f a house and tries to rape me v io le n tly .” She h its
h im and knocks h im unconscious. W h e n he starts to regain consciousness,
she escapes in a car.

In a ll, the dreams about H ow ard are a classic exam ple o f the operation
o f the re p e titio n p rin c ip le over the space o f nearly 25 years. T here is some
decline in the negative aspects o f these dreams, and there are changes in
her w aking reactions to them , b u t the issues they reflect rem ain essentially
unchanged. I t is as though the themes are embedded in her vig ila n c e -fe a r
system and are subject to re a ctiva tio n under circum stances th a t cann ot be
determ ined w ith the in fo rm a tio n th a t is available fo r th is study.
M oreover, the re p e titive patterns seem to persist even though Sanders
feels she has gained greater perspective on her fa ile d m arriage in w aking
life . H er b e lie f is supported in the interview s w ith her friends, w ho note
th a t she talks about h im less and expresses less anger th a n in th e past. B ut
the dreams do n o t change quite as m uch as her thoughts about h im . T h is fits
w ith Foulkes’s (1985) co g n itive view th a t dreams reflect m ental encodings o f
past w aking experience, n o t the experiences themselves, so they are n o t
lik e ly to be changed by fu rth e r reflections. I t also fits w ith the idea th a t
the updated aspects o f a conce ption are no m ore lik e ly to appear in dreams
th a n the older one, because o f the lack o f re a lity constraints d u rin g sleep.
T aken together, these ideas lead to the hypothesis th a t w aking feelings may
change more, or m ore q u ickly, th a n dream feelings.
The dreams about H ow ard could be said to co n ta in a “ w ish,” in the
Freudian sense, in th a t th e y clearly show a regret th a t the m arriage was n o t
successful and sometimes express a hope fo r re co n c ilia tio n as w e ll. There
is also a w ish fo r sexual g ra tifica tio n in several o f the dream reports. A t the
same tim e, the dreams have a traum atic q u a lity th a t overwhelm s the w ish fu l
dim ension. T he fears and negative em otions persist, yet p e rio dica lly, dreams
occur th a t clearly seem to “ resolve” the issue and signal th a t the dream er
is ready to move on. Those dreams are eventually follow ed by another
negative dream. T h is fin d in g should be given consideration before accepting
claim s in the c lin ic a l lite ra tu re about the frequency and im portance o f
“ resolu tion” dreams. Such claim s ofte n suffer fro m the lack o f lo n g itu d in a l
follow -up.
D ream s o f a F ailed In fa tu a tio n
In late September 1994, several years after she had dated anyone
regularly, Sanders m et a m an at a party w hom she found a ttra ctive . She
and Derek had a com m on circle o f friends and a m utual in terest in the
theater and dreams. T he y struck up a friendsh ip, and a few m onths la te r
they were in a sm all p la y w ritin g group together. Later they were in the
same dream -sharing group as w e ll. Derek is 12 years younger th a n Sanders
and d id n o t seem interested in more th a n a friendship, w h ich Sanders
basically understood. Nonetheless, she became infatuated w ith h im and
entertained the hope o f a ro m a n tic relatio nship.
H e r friends were sure th a t n o th in g w ould come o f the re la tio n sh ip
and w orried about her. However, tw o o f them fe lt th a t the re la tio n sh ip
w ould have positive aspects because it w ould add new zest to h e r life no
m atter how it ended up. T he friendship blossomed over the space o f a year.
Perhaps D erek gave some in dica tions th a t he d id care fo r her ro m a n tica lly
m ore th a n she realized a t first, m ostly through h e a rtfe lt conversations. W h a t
ever the signals, Sanders came to feel betrayed w hen th e re la tio n sh ip d id
n o t go fu rth e r. She became upset w hen he showed affe ctio n tow ard another
wom an in her presence. W h e n he came to a m eeting o f th e dream group
w ith a date, Sanders expressed her annoyance to h im , to th e great surprise
o f his date, and in effect ended the friendsh ip in early A p ril 1996. She saw
h im on ly once o r tw ice in passing after th a t.
D erek appears in 43 dream reports during the tim e period covered by
the systematic analysis, th e n in another 4 dreams in the p o rtio n o f the
dream jo u rn a l w ritte n after th is study began.3 T he firs t dream occurred o n
O ctober 7,1994, ju s t a few days after she m et h im . Thereafter, the frequency
and co n te n t o f the dream reports reflect the rise and fa ll o f her hopes about
h im . T he dreams have a h ig h rate o f frie n d ly and sexual in te ra ctio n s and
a lo w rate o f aggressions, especially physical aggressions, as show n in the h
p ro file in Figure 5.3. T h is h profile compares dreams concerning D erek to
dreams w ith H ow ard in them and to the baseline sample (D u n n , 2000).
T h irte e n o f the firs t 16 dream reports in w h ic h D erek appears co n ta in
sexual or in tim a te physical interactions w ith h im , such as w arm hugs o r
resting her head on his lap. These early sensual dreams are in general very
positive and fu ll o f a n ticip a tio n , b u t they also express her fear th a t he does
n o t care about her; in one dream, he even chases her after he has an orgasm:
November 11, 1994: Derek kisses me. I am disappointed because his

mouth is hugely wide and it doesn’t feel good. He French kisses and it
is intense, our tongues intricately intertw ining. I feel him have an
orgasm and I am untouched w ith sexual passion. It feels very cold and
self-serving. Now I am being chased by him and other people. I run for
my life.
As Sanders accepted th a t the relatio nship was n o t going to develop

in the way she hoped, her sexual interactions w ith D erek become less
frequent in her dreams. She also has dreams in w h ich she is angry w ith h im
o r jealous th a t he is h aving sex w ith someone else. He is m ore peripheral
in the dream reports in w h ich he does appear, and some characters are now
described n o t as “ D erek,” b u t as “ lik e D erek.” O n M ay 2 and 3, 1996, about
3O nce again, as w ith the H ow ard subseries, this part o f the analysis includes dream reports from after
the firs t 3,116 dream reports were received in order to make the subseries even more useful.
123
- 0.6
Befriender %
Aggressor %
AggresslorVChsracter Index
Friendllness/Character Index
Sexuallty/Character Index
At Least One Aggression
A t Least One Friendliness
A t Least One Sexuality
Figure 5.3. An h profile of B arb S anders’s dream s o f “D erek” and “ H ow ard,” com pared
w ith th e Barb S anders baseline 250 sam ple.
a m o n th after she broke o ff th e ir friendship, Sanders describes dreams in

w h ic h she is upset because he is w ith another woman:
I am sobbing. M y heart is broken. Derek has betrayed our relationship

w ith another woman. N ow he follows me around begging me to forgive

h im . I can’t stop crying. D erek is w ith someone else or is ign o rin g me.
I am s ittin g n e x t to his b ro th e r and decide to h e ll w ith D erek, I w ill
chase his broth e r. W e hug and flir t. D erek is jealous and moves between
us, and now I am havin g d in n e r w ith D erek and th e n a h e lico p te r
comes and picks me up in the m iddle o f din n e r.
Then, after a period o f 5 m onths in w h ich she records 72 dreams, b u t

none o f Derek, Sanders has one th a t she calls “ the kiss o f forgiveness” :
September 26, 1996: I am s ittin g on a couch. A m ale frie n d w ho is

also a frie n d o f D erek’s comes in , and D erek is w alking behind h im . I
am surprised. H e comes in and talks a b it. I respond back in a frie n d ly
m anner. H e is s till u p tig h t about our break-up and n o n -co n n e ctio n o f
the last few m o n th . T he frie n d comes over and sits beside me on the
couch. D erek walks over and puts his face betw een our faces. I kiss h im
on the cheek. H e is s till distant, b u t then he keeps ta lk in g and I keep
ta lk in g and th e n he kisses me on the lips in a genuine kiss.
A week later, she has a dream th a t im plies they are s till friends, even
though she has n o t seen h im in m onths:
October 3, 1996: D erek approaches me and begins to ta lk quite

frie n d ly and I respond back. I t ’s lik e we are pretending the last few
m onths o f no com m unication never happened. I feel glad we are ta lk
ing again.
A fte r th is p a ir o f dreams, D erek appears in o n ly 4 o u t o f the n e x t 461

dream reports. By contrast, he appeared in 43 o f the previous 334 dream
reports. Tw o o f the dreams after the apparent resolu tion dreams are re la tive ly
benign, b u t in one someone lik e Derek is betraying her, and in the o ther
he is torm entin g her, so he is n o t e n tire ly gone from her m ind:
February 16, 1998: I am try in g to do som ething. T o get som ething

done. D erek interferes, teasing me and stopping me fro m accom plishing
w hat I am try in g to do. I am very angry a t h im , b u t I choose n o t to
blow up. I h o ld in m y anger because I know he is deliberately try in g
to provoke me. H e is a cting lik e one o f m y granddaughters does. I am
try in g to be lo v in g in spite o f th e provocation .
T he Derek dreams are s trik in g fo r the fa ct o f sexual intercourse and

o th e r sexual intim acies th a t have no correspondence to her w aking re a lity.
T he y are clearly w ish fu l dreams: They are continuous w ith her w aking
hopes, but n o t w ith her w aking life . T he dreams also d ire ctly reflect her
fears about h im and the relationship. U n lik e the H oward subseries, they
decline greatly in frequency once the wishes are gone. T h is p a tte rn leads
to a testable hypothesis fo r future studies: I t is o n ly possible fo r a sig n ifica n t
person to disappear from dream life if no re a l-life in tim a te in teractio ns have
been involved.
NEW WAYS T O STUDY MEANING IN DREAMS 125

For several reasons, the D erek subseries is deserving o f m uch more
thorough study th a n has been possible here. I t is long enough to make
detailed analyses possible, b u t n o t so long as to be overw helm ing. I t occurs
in a re la tiv e ly circum scribed period o f tim e and is m ixed w ith 334 o th e r
dream reports between O ctober 1994 and O ctober 1996. I t has been coded
fo r several H a ll-V a n de C astle categories, the results o f w h ich are available
in the SearchCodings section ofD ream R esearch.net. In a d d itio n , it contains
m uch m ate rial th a t m ig h t lend its e lf to a b lin d m etaphoric analysis th a t
could be com pared w ith the coding results. F inally, the in tervie w s w ith
Sanders and a ll fo u r friends provide h ig h ly detailed com m entary on the
relatio nship. M ore generally, it is possible th a t h ig h ly focused analyses o f
subseries such as th is one may be an excelle nt way to study dream m eaning
in great de ta il.
For now , though, enough analysis has taken place w ith the H ow ard
and Derek subseries and the o ther character studies presented in th is section
to show th a t Barb Sanders’s dream reports present an accurate p o rtra ya l o f
her relationships w ith the im p o rta n t people in her life . T he findings support
a hypothesis th a t derives fro m the studies o f the L u c ille series and m any
other dream journals: People’s dreams about fa m ily and friends enact th e ir
conceptions and concerns in regard to them . Such dreams are lik e dramas
th a t reflect w aking relationships; they are m in ia tu re soap operas.
A lth o u g h th is is an im p o rta n t and useful conclusion, it does n o t fo llo w
th a t a ll the o ther elem ents in these dreams are equally accurate and in fo rm a
tive . Nonsense may be m ixed in w ith the coherent aspects in dreams about
fa m ily members and close friends. N o r does it fo llo w fro m these findings
th a t the m any dreams th a t do n o t include know n characters are consistent
w ith w aking conceptions. Thus, a great deal o f dream co n te n t rem ains to
be explained.
T he ate r Dream s
I t appears from reading through Barb Sander’s dreams th a t she has a

strong interest in th e a trica l perform ances, w hether as an actor, singer, or
director. W h e n a w ord string co n ta in in g a ll relevant theater-related term s
is entered in to D ream Bank.net, it retrieves 169 dream reports fro m the
series, 77 o f w h ic h actually in vo lve Sanders a u d itio n in g for, ta kin g p a rt in ,
o r d ire ctin g a th e a trica l p rodu ction. T h is large num ber o f theater dreams
is, in fact, continuous w ith Sanders’s w aking interests. She acts and sings
in productions, some o f w h ich she w rites herself. She also enjoys d ire ctin g
th e a trica l perform ances.
In h a lf o f the theater dreams in w h ich she is in vo lve d as a perform er
o r d ire cto r, Sanders sees herself as g ivin g an excellent perform ance. Some-

tim es she comes out o f the audience to give an unexpected perform ance
th a t is m et w ith great approval. In others she is “ discovered” at the last
m in u te to be w o rth y o f a starring role. I t is clear from these positive theater
dreams th a t Sanders has, or once had, h ig h hopes and am bitions. She w ould
lik e to be in the p u b lic lim e lig h t as an esteemed a rtistic figure. H e r desire
is shown m ost dram atically in a dream report in w h ich she, her m iddle
daughter, and another woman fin d themselves singing on stage to great
acclaim . Sanders reports th a t she loves the applause b u t wants to do a solo,
and she ushers her daughter and the o ther wom an aside:
June 8, 1985: I see a stage, w ith m usicians there. I go o u t on stage,

to the m ike. I remember th a t m y sister and daughters are there too. I
start to ta lk to the audience, w ho are fa r away, seated a t tables. I say,
“ everybody get up and go to the dance flo o r.” M ost do. . . . I tu rn to
the m usicians, “ H it it boys, some h o t boogie. . . . ” I start singing and
i t ’s good. M y m iddle daughter and another w om an jo in in , three-part
harm ony. I t ’s great. A udience loves it, I love, it except I w ant to do it
solo. I te ll m y daughter and the o th e r wom an to go away and I sing
ballad. I t ’s great.
T his inference o f great a m b itio n is supported in the in te rvie w w ith

her frie n d Lucy, w ho often perform s w ith her:
I th in k she’s in cre d ib ly am bitious and overachieving, and i t ’s lik e it ’s

never enough. N o th in g is ever enough fo r her to feel good enough
about herself, and yet she’s so wise and good and strong and able, and
it ’s ju st never enough fo r her to m ake herself feel, I th in k th a t’s w hy
she goes o u t and does theater because she wants to make herself feel
positive and beloved.
But an equal num ber o f the theater dreams co n ta in rejections and

m isfortunes. Sanders does n o t w in the part, or people leave as she is about
to perform . She misses a rehearsal, can’t fin d the theater, or nearly falls o ff
the stage. In seven instances, she forgets her lines or is afraid she w ill forget
them , and in tw o others she does n o t have a script. T h e negative events
are consistent w ith tw o o f her w aking concerns in regard to her public
performances. First, Sanders is indeed afraid she w ill forget her lines, as
attested to by b o th her and her friends. Second, she does feel th a t she is
o fte n ignored o r unappreciated, a p o in t th a t is stated m ost fra n k ly by tw o
o f h e r friends.
So, ju st as Sanders’s dreams featuring sig nifica nt others enact her
conceptions o f her relationships w ith them , the theater dreams seem to be
variations on a few o f her m ajor concerns about a rtis tic perform ances. She
wants to be notice d and th in ks she deserves far more a tte n tio n th a n she
receives, b u t she worries th a t she w ill forget her lines and th a t people w ill
127
ignore her. These concerns are the “ themes” th a t are acted out to various
degrees in each dream re la tin g to the theater and perform ances. T h a t is,
eaclrdream can beseen as a specific instance fro m w h ic h generic in fo rm a tio n
can be extracted. T h is in s ta n tia tio n o f generic in fo rm a tio n may be based
o n an abstract conceptual m etaphor, “ the generic is specific” (Lakofif, 1993a;
L a ko ff & T urner, 1989). T he existence o f th is m etaphor is inferred from
the fact th a t people can understand parables so readily, in clu d in g parables
fro m o th e r cultures th a t they have n o t heard before. I f the “ generic is
specific” m etaphor could be shown to be operative in generating at least
ce rta in types o f dreams, such as ones th a t are variations on a them e, it
m ig h t account fo r a sig n ifica n t m in o rity o f dream reports.
D ream E lem ents T h a t A re N o t C o n tin u o u s
A lth o u g h the several analyses presented so far support the c o n tin u ity
p rin cip le , some elements in Sanders’s dream reports are n o t continuous w ith
her w aking life . They may provide an in terestin g exception to the c o n tin u ity
p rin c ip le th a t could lead to a better understanding o f dream m eaning, o r
perhaps they reveal the lim its o f the conceptual systems available to the
neural n e tw o rk fo r dream ing (Foulkes, 1999; H obson, 1988). For exam ple,
Sanders has several dreams about cats, especially kitte n s, th a t are neglected,
deform ed, o r starving. T he appearance o f cats fits w ith her interest in cats
in w aking life , but contrary to the c o n tin u ity p rin c ip le , she does n o t w o rry
about the h e a lth o f cats in w aking life , n o r does she fear th a t they m ig h t
starve o r be neglected.
A n exam ple o f how the co n tra d icto ry findings on neglected cats m ig h t
be approached using Dream Bank can be seen through the m ention o f “ stray
k itte n s .” Five o f the eig h t uses o f the term “ stray” occur in c o n ju n c tio n
w ith kitte n s, a contingency w ith a p value o f .000. Four o f these instances—
in January 1981, O ctober 1982, and tw ice in O ctober 1986— occur in
reference to m en w ho are lost souls w ho do n o t am ount to m uch in her
eyes. T he fifth , w h ich occurred in December 1981, concerns tw o actual
stray kitte n s. T he dream report begins as follow s, th e n moves to unrelated
topics having n o th in g to do w ith kitte n s o r inadequate men:
T h e stray k itte n s plead fo r food. T hey are very hungry. I feel badly fo r
them . I lo o k in th e refrigerator. I fin d some sugar cakes and some cheese.
I am dressing up to go o u t on a date. I t ’s a conservative o u tfit, b u t as
soon as I ’m o u t th e door, I ’ll readjust the fro n t and it w ill be very sexy.
In w aking life , the equation o f “stray k itte n s ” and “ lo st m en” is under

standable to m ost people, given th e ir general know ledge o f the w orld, as a
conceptual blend. T he y see the conn ection due to th e ir instantaneous
com prehension o f the characteristics o f stray k itte n s th a t can be applied to

at least some m en— helplessness and lack o f attachm ent to an im p o rta n t
source— w h ile ignoring the kitte n s ’ irre le va n t properties— needing to be
breast fed, furriness, claws, and m ew ing (Teenie M a tlo ck, D epartm ent o f
Psychology, Stanford U n ive rsity, personal com m unication, A ugust 15,
2001). W ith regard to Sanders’s dreams, the research question is: Does h e r
understanding o f rootless men as stray kitte n s underlie h e r dream o f actual
stray kittens? T h is exam ple m erely poses the question and in no way begins
to answer it.
U n lik e kitte n s, horses are portrayed in a positive lig h t in Sanders’s
dream reports. She has 24 dreams in the series in w h ich she is rid in g a
horse. In a ll o f those dreams, she portrays herself as an excellent rider. In
one dream she rides lik e the w ind, in another she learns q u ickly and becomes
an excelle nt rider, and in another her fath er praises her fo r her rid in g skills.
Taken together, the dreams give the im pression th a t she learned to ride as
a c h ild and likes horses, but th a t is n o t the case. The fo llo w in g excerpts
from the in te rvie w w ith her give the flavor o f her reactions to inferences
about her conceptions and w aking experiences concerning horses:
When you were growing up, did you learn to ride horses? I had some
experiences w ith rid in g horses, yes, b u t I d id n ’t have a lo t o f it. I loved
doing it. T hey scared me; they were a w fully big and they had a tendency
to b ite . B u t other th a n th a t, i t was fun. I do rem em ber a couple o f
incidents where we were galloping away, and it was such, o u t o f c o n tro l,
b u t fu n k in d of, you know , I was ju s t hanging on fo r dear life , hoping
I don’t fa ll. But you wouldn’t say you did a lot of riding or were a good
rider? N o , I was n o t a good rid e r.
But you’re a good rider in your dreams. I am, aren’t I, yes. Did you
think they are pretty positive dreams, your horse dreams? Yes, I th in k m ost
o f my anim al dreams seem to be re a lly p ositive dreams, yes. But yet, in
this case we’re seeing then an example of where your proficiency in riding
and your use of horses and all really doesn’t reflect reality. I t ’s very m etaphor-
ical. I t ’s n o t, it ’s n o t a real life experience. I th in k I have ridden horses
m ax e ig h t tim es in m y w hole life , you know , and I can rem em ber the
specific tim es. Somebody else h e ld the horse, one had a tendency to
b ite and I was very afraid o f h im , and we sort o f w alked slow ly around
the pasture. A n o th e r one, th a t one was okay, b u t it got spooked by
another horse th a t started ru n n in g , so I was ju st hanging on fo r dear
life . A n d w hen they d id tro ttin g , it was ju st disaster c ity because I never
learned how to do th e posting.
A lth o u g h the dreamer begins by saying she “ loved” the few occasions
on w h ich she rode a horse as a c h ild , the general th ru st o f the in te rvie w
contradicts the expectations o f th e c o n tin u ity p rin cip le . She la te r says th a t
horse dreams are “n o t re a lity .” T h e p o in t could be stretched to say th a t the
dreams are continuous w ith her wishes, b u t the problem is th a t the c o n tin u ity
NEW WAYS T O STUDY MEANING IN DREAMS 129

p rin c ip le w ould p redict th a t her negative experiences w ith horses should
be reflected in the dreams as w e ll. Later in the in te rvie w , she recalled m ore
about horses as follow s:
One more thing about horses— my father became a part owner o f a

race horse, and for a while this beautiful race horse was in our back
pasture, and so it wasn’t like an animal I could ride or anything, but
it was quite exquisite and I would like to look at it as it ran around,
and we would go to the races and cheer for the horse. Yeah, and he
was beautiful, and that was a special arena, horse racing, so he was
supposed to go out there and make money, [she laughs]
B e a u tifu l horses are part o f Sanders’s pleasant memories and a plausible

basis fo r her p o sitive dreams o f rid in g horses. T he horse dreams seem to be
based on her early thoughts about horses w ith little regard fo r her few and
o fte n negative w aking experiences. Thus, the research issue is how to te ll
these possible w ish dreams from re a listic dreams.
Sanders’s dreams also show a la ck o f c o n tin u ity in re la tio n to her use
o f guns and rifles in her dreams. In 33 dream reports in the series, she is
h o ld in g o r shooting a gun, always w ith confidence. She fixes and reloads
guns, k ills dangerous anim als, and fights o ff hum an attackers. In one dream
she captures seven m en at gun point. In another she grabs a w om an’s gun
and k ills her. Based on the c o n tin u ity p rin c ip le , these assertive dream actions
led to the inference th a t she m ig h t have learned to shoot guns as a c h ild
and s till enjoys doing so. H ow ever, the inference is in correct, as seen in
the fo llo w in g in te rvie w m aterial:
H ow about guns? Were there guns around? Don’t like guns. There weren’t
a lo t o f guns around. Now in extended fam ily— uncles and cousins and
things, you know, a bunch of rednecks and guns are part o f that culture—
but in my family, no. B ut as fa r as you shooting them and shooting w ith
them. . . . I did some target work when Howard and I were first starting,
because he was interested in guns. He had guns all the time and loved
hunting. D id you ever do any hunting w ith him? He wanted me to, and
he took me out target shooting, but I balked at actually shooting animals.
I have the Bambi syndrome. D id you do much target shooting? D id you
feel proficient w ith a gun? Umm, no, not a lot.
T he unexpected answers to questions about horses and guns may p ro

vid e a sta rtin g p o in t fo r future studies: Dream co n te n t th a t is n o t continuous
w ith w aking memories and past experiences may be in d ic a tiv e o f figurative
though t. In e xploring such a p o ssib ility in the Barb Sanders series, it w ould
have been useful to start w ith questions re la tin g to her thoughts and fantasies
concerning horses and guns. T h a t is, the focus should have been on how
she organizes and uses her know ledge and feelings about horses and guns
before tu rn in g to her actual experiences w ith them . T his conclusion could
be o f use in future investigations o f figurative thought in dreams
U n u su a l Elem ents in D ream R eports
T he unusual elem ents in dreams— distortions in fa m ilia r settings, im

possible acts like fly in g under one’s ow n power, and metamorphoses—
im m ediately come to m in d w hen m ost people in W estern cultures th in k o f
dreams. A lth o u g h such elements are less frequent in representative samples
o f dream reports th a n popular stereotypes suggest, they do happen in 10%
to 35% o f dreams (H a ll, 1966b; Revonsuo & S a lm iva lli, 1995; R ittenhouse,
S tickgold, & Hobson, 1994), and they once again raise the p o ssib ility th a t
dreams have nonsensical aspects resulting from the lim ite d capabilities o f
the conceptual systems available during sleep.
I t also may be th a t some unusual elem ents are nonsensical and others
are m etaphorical, m eaning th a t d istin ctio n s among various types o f unusual
elements m ight be useful. For example, com posite characters th a t are based
on tw o different people sometimes appear in the Sanders dream reports,
usually designated by a slash, as in “D w ig ht/H o w ard,” a com posite o f her
fa vo rite brother and ex-husband. C om posite characters lend themselves to
study as possible conceptual blends (G rady et ah, 1999). In the 43 dreams
in w h ich Derek appears, fo r example, 8 o f those appearances characterize
h im as “ lik e ” someone else or as a com bin ation o f him self and another
person. M ost strik in g ly , he appears as “D e rek/D a rryl” in three dream reports,
w h ich yokes h im w ith the true love th a t she rejected fo r unfaithfulness at
age 18. In this case, there are tw o pieces o f in fo rm a tio n th a t suggest th a t
this com posite character may represent a conceptual blend. F irst, Sanders
said in the in te rvie w th a t they were the tw o m en she loved th e m ost.
Second, they are also sim ilar in th a t she never had sexual intercourse w ith
e ith e r o f them .
Interestingly, 5 o f D a rryl’s o ther 31 appearances in the dream series
in vo lve comparisons or composites in a d d itio n to the three composites w ith
Derek. T he netw ork o f likenesses and composites centering on D erek and
D a rryl is presented in Figure 5.4. I t is another example o f how the study
o f unusual elements in dreams m ig h t proceed.
C haracter metamorphoses are re la tive ly rare in dreams; they occur
only 12 tim es in the 1,000 dream reports com prising the H a ll-V a n de Castle
norm ative samples. H ow ever, they ofte n strike dreamers as rem arkable and
m ysterious when they occur. A lth o u g h it is com m on in w aking life to say
th a t a person is lik e someone or fo r an object or event to rem ind someone
o f some other object or event, metamorphoses nonetheless seem far rem oved
from w aking though t patterns, even though they are often seen in m ovies

Figure 5.4. C om posite characters involving “D erek” and “ D arryl” in B arb S a n d e rs’s
dream s. Each line represents an instance in w hich tw o characters w ere “c o m b in e d .”
and videos. T hey therefore provide an in terestin g challenge and o p p o rtu n ity
fo r researchers w ho hope to fin d m eaning in a ll aspects o f dream content.
T here are o n ly fo u r instances o f character metamorphoses in the Sand
ers baseline 250, n o t enough fo r a system atic study. T o fin d a sample o f
metamorphoses in the e n tire Sanders series, the terms “ changes in to ,” “ turns
in to ,” “ becomes,” and “ is now ” were entered in to D ream Bank.net. I t is
u n lik e ly th a t the fo u r terms capture a ll the metamorphoses in th e series,
b u t they do provide a large sample th a t is probably representative o f the
p o p u la tio n o f m etam orphic changes in the series. T he sample has the added
advantage o f in c lu d in g metamorphoses o f objects, w h ich are n o t coded fo r
in the H a ll-V a n de Castle system. A fte r e lim in a tin g phrases such as “ the
argum ent turns in to a fig h t,” “he becomes angry,” and “ the food is now
ready to serve,” th e in itia l yie ld o f 132 dream reports boiled dow n to 50
instances th a t q u a lify as metamorphoses in 49 dream reports.
T h irty o f the 50 metamorphoses include a hum an o r anim al character
at the beginning o r the end o f the transform ation. T h irte e n o f those changes
are hum an-to-hum an transform ations, b u t in seven dream reports people
change in to anim als, creatures, or objects; in another seven, the anim als
o r objects tu rn in to people. O n tw o occasions, one anim al turns in to another,
and in one case, a m ale puppet turns in to a fem ale puppet. Thus, character
transform ations in v o lv e no one pattern .
A few o f the changes seem to be sim ila r to the cases o f com posite
characters discussed in the previous section, such as w hen Sanders’s ex-
husband is now one o f her brothers, o r a m an is now lik e Derek, o r a wom an
turns in to Faye Dunaway. I t is also in terestin g th a t n in e cases in vo lve
babies o r young ch ild re n , w ho are obvious instances o f re la tiv e ly rapid

transform ation. However, there are several changes th a t do n o t seem to
have any im m ediate p la u sib ility, even when the co n te xt o f the dream is
added. For example, a yellow wooden horse turns in to an a rtis tic man; a
horse becomes a cat; a large pig becomes a pigle t; and a spider becomes a
m iniatu re m an, who in tu rn becomes a lig h t bulb.
In a d d itio n to the 30 transform ations th a t in vo lve dream characters,
there are 20 cases where one object turns in to another. Fourteen o f those
changes in vo lve objects th a t are lin k e d to trave l or m ovem ent, a fin d in g
th a t may suggest a possible underlying pattern connected to one o f several
conceptual m etaphors. For example, “ change is m o tio n ,” as in “ his m ood
jum ped from morose to ecstatic in less th a n a m in u te ,” and “processes are
m ovem ents,” as in “ the m eeting goes for tw o hours” (T eenie M atlock,
D epartm ent o f Psychology, S tanford U n ive rsity, personal com m unication,
A ugust 15, 2001). B o th m etaphors are pervasive in everyday th in k in g , and
they are com patible w ith a m etaphor closely related to issues o f self: “ life
is a journey,” as in “ she’s come to a fo rk in the road” and “ they have one
more m ountain to clim b .” A cco rd in g to L a ko ff (1987), these and other
m o tio n m etaphors are connected to an “ um brella” (i.e., higher le ve l) concep
tu a l m etaphor called event structure.
Regarding the specific o b je ct-to -o b je ct metamorphoses in the Sanders
series, several o f the transform ations are fa irly straightforw ard, such as when
a sm all car becomes a big, flashy car; a car becomes a house tra ile r; or a car
becomes a hearse. O th e r changes in v o lv in g vehicles seem less straightfo r
ward, as w hen a bus becomes a series o f kid die cars, a table becomes a circus
tra in , or a pickup tru c k becomes a tape recorder th a t has to be pushed up
a h ill. In a d d itio n , interesting changes th a t may relate to “ change is m o tio n ”
or “ life is a journey” occur in passageways: A bridge becomes a stairway,
another bridge becomes a boat, a road becomes a hallw ay, another road
becomes a staircase, a w a terfall becomes a freeway, and a hallw ay is suddenly
the inside o f a diesel truck.
A lth o u g h the 50 cases suggest some possible m etaphoric meanings, it
was n o t possible to develop con vin cin g evidence th a t any o f the m etam or
phoses relate to the systematic findings presented earlier in th is chapter,
despite the co n te xt provided by the series. I t w ill take fu rth e r investigations
and, perhaps, new approaches, if a case is to be made fo r m eaning in
these elements.
C O N C L U S IO N
T his chapter has o n ly scratched the surface o f the Barb Sanders series.
I t is a dem onstration o f possibilities, n o t a d e fin itiv e analysis. However,
enough has been said to illu stra te th a t the m ain characters, social interac-
N E W WAYS T O STU D Y M E A N IN G IN DREAMS 133

tions, and a ctivitie s in the dreams reveal Sanders’s w aking conceptions and
concerns in re la tio n to the sig n ifica n t people and interests in h e r life . O n
the o th e r hand, there are elem ents in the dreams th a t do n o t im m ediately
seem continuous w ith her w aking conceptions, such as her e xce lle n t rid in g
and shooting. Those elem ents may reveal the lim its o f co g n itive capabilities
during sleep, or they may be the products o f figurative th in k in g . S im ila rly,
the unusual elements in the dream reports, such as com posite characters
and metamorphoses, may define a dim ension th a t goes fro m the m etaphorical
to the nonsensical. T he reso lu tio n o f these seeming anom alies w ill require
m any fu rth e r studies.
M ost im p o rta n t, th is chapter illustrates a m ethodology th a t could be
applied to a w ide range o f in terestin g cases. Perhaps it w ould be especially
revealing w ith people w ho were keeping a dream jo u rn a l before they suffered
a b ra in in ju ry or a personal traum a or had to go on one o r another m edication.
I f such people were th e n able to resume th e ir dream journals, the o rig in a l
jo u rn a l could serve as a baseline fo r try in g to determ ine w h at effects, if any,
the sudden a lte ra tio n in th e ir life circum stances is having on th e ir dreams.
T he new findings in th is chapter, along w ith the m any e m p irica l studies
cite d in previous chapters, provide a basis fo r futu re advances in the study
o f dream m eaning. T hey also make it possible to develop an assessment o f
tra d itio n a l dream theories, w h ic h is the m ain task o f the fin a l chapter.

6
A CRITIQUE OF TRADITIONAL
DREAM THEORIES
T his chapter brings together a ll systematic findings on dreams, in c lu d

in g some o f the new m aterial presented in earlier chapters, to address the
m ain hypotheses in the best know n tra d itio n a l theories o f dreams, and
concludes th a t a ll o f these theories have failed on one or another key issue,
suggesting th a t it is tim e to fin d a new starting p o in t. T o begin, the chapter
discusses every m ajor cla im made by Freud and Jung, none o f w h ich stands
up against the em pirical evidence th a t is available. H ow ever, the chapter
does n o t consider the c lin ic a l theories o f the neo-Freudians, phenom enolo-
gists, or existentialists because th e ir theories are e ith e r com binations o f
Freudian and Jungian ideas to one degree or another or deductions th a t are
based on a philo sophical tra d itio n . T he neo-Freudians A d le r and From m , fo r
exam ple, drew on M arxism to make general statem ents about the c o n tin u ity
between dream ing and w aking though t th a t tu rn o u t to have v a lid ity in
the lig h t o f la te r research, b u t th e ir w ritin g inspired no systematic w ork on
dreams and had no lasting im pact.
F ollow ing the discussions o f Freudian and Jungian dream theories, the
chapter addresses the problem s w ith both the o rig in a l and revised versions
o f a ctiva tio n -syn th e sis theory. T h is theory has the m ost com m onalities w ith
the neurocognitive m odel presented in this book, b u t sig n ifica n t differences
rem ain. F in a lly, the chapter points o u t the weaknesses o f the m any theories
th a t assume one o r another problem -solving fu n c tio n fo r dreams. Some
o f those theories hypothesize th a t recalled dreams have problem -solving
functions, whereas others cla im th a t the functions are carried o u t during
sleep, w hether the dreams are recalled or n o t. E ith e r way, problem -solving
theories have been extrem ely d iffic u lt to test, so they rest p rim a rily on
analogies w ith new discoveries in other fields or on anecdotal examples.
The evidence used in assessing the theories comes from a w ide range
o f areas w ith in em pirical dream research, in clu d in g experim ental studies in
sleep laboratories, co n te n t analyses o f dream reports, neuropsychological
studies o f the effects o f b ra in lesions on dream ing, and co rre la tio n a l studies
re la tin g dream re ca ll to co g n itive and personality variables.
135
T H E F R E U D IA N T H E O R Y O F D R E A M S
D u rin g the la te 19th century, dream theorists generally believed th a t

dreams are b rie f and th a t they are usually a reaction to an in te rn a l or
external stim ulus or occur during the process o f awakening. Freud (1900)
trie d to blend these perspectives by com paring dreams to “ a fire w o rk th a t
has been hours in the preparation, and th e n blazes up in a m om ent” (p.
377). H e agreed th a t they last fo r o n ly a b rie f tim e and perhaps occur on ly
during awakening, b u t he added the new idea th a t the thoughts underlying
dreams develop slow ly during the day. H ow ever, contrary to Freud, laboratory
studies have revealed th a t dream ing takes place longer, m ore frequently,
and more regularly th a n he or any o ther theorist ever im agined before the
serendipitous discovery o f R EM sleep in 1953 (A serinsky & K le itm a n , 1953;
Dem ent, 1955; D em ent & K le itm a n , 1957b).
Freud (1900) also asserted th a t “ a reference to the events o f the day
ju st past is to be discovered in every dream ” (p. 127), but five detailed
studies la te r dem onstrated th a t o n ly about h a lf o f dreams co n ta in even the
slightest “ day residue” th a t can be id e n tifie d by the dreamer (B otm an &
C rovitz, 1989; H a rlo w & R o ll, 1992; H artm ann, 1968; M arquardt, Bonato,
& H offm ann, 1996; N ielsen & Pow ell, 1992). A s part o f his emphasis on
th e large role o f specific memories in shaping dream co n te n t, Freud (pp.
2 6 6 -2 6 7 ) concluded th a t a ll sig n ifica n t speeches in dreams can be traced
to memories o f speeches heard or sentences read, but the analysis o f hundreds
o f speech acts in dream reports collected in sleep laboratories has shown
th a t they are usually new constructions appropriate to the u n fo ld in g dream
context, n o t reproductions (M eier, 1993). Indeed, speech acts are sometimes
so appropriate to the dream co n te xt th a t m any b ilin g u a l pa rticip a n ts in one
laboratory study reported th a t they spoke in the language understood by
the dream character w ith whom they were ta lk in g (Foulkes, M eier, Strauch,
Kerr, Bradley, & H o llifie ld , 1993).
Freud’s (1900) m ost famous and im p o rta n t claim is th a t “ w is h -fu lffll-
m ent is the m eaning o f each and every dream ” (p. 106). A lth o u g h th is
hypothesis is based on his w ork w ith adult patients, he th o u g h t th a t the
dreams o f young c h ild re n provide “ invaluable p ro o f’ o f his w is h -fu lfillm e n t
theory, and he used dreams from his ow n ch ild re n as evidence, in clu d in g
a sleep-talking episode from his 19-m o n th -o ld daughter. H ow ever, Foulkes
(1982, 1999) found th a t young c h ild re n have static and bland dreams, n o t
a t a ll lik e Freud’s anecdotal examples, and concluded th a t there are no signs
o f wishes in c h ild re n ’s co g n itive ly im poverished dream reports. M oreover,
examples from sleep-talking episodes no longer have any standing as e v i
dence since the discovery th a t m ost sleep ta lkin g originates durin g the
m icroaw akenings o f from 10 to 20 seconds th a t occur several tim es per
n ig h t in b o th c h ild re n and adults (A rk in , 1981; Boselli, Parrino, Sm erieri,
&. Terzano, 1998; M a th u r & Douglas, 1995)
Freud began the case fo r his w is h -fu lfillm e n t theory w ith c h ild re n ’s
dreams because he believed th a t the wishes in most adult dreams are disguised
in order to reduce th e ir anxiety-arousing tendencies. T h is self-deception is
especially the case fo r the in fa n tile sexual wishes th a t Freud believed are
the basis fo r m ost a d u lt dreams. Dreams are disguised at the behest o f a
“ censor” by four co g n itive processes th a t together comprise the dream-work
(1900, pp. 328, 389). Displacement is a process whereby h ig h ly charged
thoughts are transferred to m in o r elements in the im pending dream. Conden
sation compresses several d iffe re n t dream-thoughts (i.e., the em bodim ents o f
wishes) in to a com posite elem ent, in clu d in g dream characters who have
the qualities o f tw o o r more people. T he tw o processes produce m ost o f the
transform ations th a t render dreams d iffic u lt to understand.
A lth o u g h displacem ent and condensation are “ the tw o forem en in
charge o f the dream -w ork” (Freud, 1900,p. 235), they are jo in e d by tw o other
processes th a t make the dream more coherent. T he regard fo r representability
changes abstract thoughts in to a p ic to ria l form th a t is m ore sensible and
acceptable to the censor. T his process can make use o f the w aking processes
o f figurative though t th a t generate jokes, legends, and proverbs (Freud, 1900,
p. 259). F inally, the dream is shaped by secondary revision, a process th a t
overlaps w ith w aking though t and is responsible fo r the in te rp o la tio n s and
additions th a t give the m anifest co n te n t a somewhat in te llig ib le pattern.
Freud (1900) brought fo rth numerous c lin ic a l examples to dem onstrate
how each o f these processes works. H e concluded by stressing th a t “ the task
o f transform ing the unconscious thoughts in to the dream -content is peculiar
to the life o f dreams” and th a t th is dream -w ork is “ far m ore rem ote from
the m odel o f w aking thought than even the most determ ined b e little rs o f
the psyche’s feats o f dream -form ation have though t” (pp. 3 2 8 -3 2 9 ). N o
con vin cin g n o n c lin ic a l studies, how ever, have dem onstrated the operation
o f the dream -work. C ontrary to Freud’s emphasis on the puzzling nature o f
dreams, tw o sym pathetic reviews o f a ll available experim ental and correla
tio n a l studies concluded th a t m ore in fo rm a tio n is available in the m anifest
content o f a dream th a n w ould be expected if the dream -w ork had made
the m anifest dream -thoughts re la tive ly meaningless (Fisher & Greenberg,
1977, 1996). T h is is also the conclusion o f a careful study in w h ich dreams
were collected from tw o participants in the laboratory and th e n compared
w ith m aterial from th e ir psychoanalytic sessions and structured interview s
(Greenberg, Katz, Schwartz, & Pearlman, 1992).
T he m ost consistent defense o f Freud’s ideas about the fo rm a tio n o f
dreams is based on an appeal to evidence from sublim inal stim u la tio n studies
claim ing th a t unconscious processes can be influenced sig n ifica n tly through
A CRITIQUE OF TRADITIONAL D R EAM THEORIES 137

the presentation o f psychodynam ically releva nt s tim u li (S hevrin , 1986,
1996; S lipp, 2000). In several studies between 1918 and 1960, particip ants
were exposed to b rie f presentations o f visual o r aud itory s tim u li below the
threshold o f conscious awareness, and th e n asked to report th e ir dreams the
n e xt m orning and draw pictures th a t were based on those dreams. T he
evidence fo r incorporations in these studies was based on seeming physical
resemblances and o n sym bolic in terpretations o f the k in d the findings were
m eant to support. T h e m ethod was therefore n o t independent o f the theory.
In a study by Fisher (1954), fo r example, it was necessary to believe, in the
face o f a ll the evidence show ing the ra rity o f stim ulus in co rp o ra tio n in to
dreams, th a t a vague draw ing by the p a rtic ip a n t had some conn ection to a
picture th a t was shown to her fo r less th a n a second below the threshold
o f awareness.
T he m ost te llin g criticism s o f the early studies have been provided by
researchers trained in the tra d itio n o f su b lim in a l psychodynam ic studies by
one o f the o rig in a l investigators, Charles Fisher. T hey noted th a t “ suitable
controls were ofte n la cking , and the various c lin ic a l in terpretations could
be seen as equivocal and, at tim es, a rb itra ry” (S hevrin , 1996, p. 96). T he
same investigators cla im th a t th e ir ow n studies introduce the necessary
controls, b u t th e ir studies re ly p rim a rily on the p ro d u ctio n o f w aking m ental
images and free associations after laboratory awakenings from R E M and
N R E M sleep, so they have n o th in g to do w ith dreams d ire ctly. In one la te r
study, participants were exposed su b lim in a lly to a picture o f a w ritin g pen
placed n e xt to a person’s knee, w h ich presum ably prim es fo r u n lik e ly cogni-
tiv e associations lik e “ pen-ny” through the clang association o f “pen” and
“ knee.” T hey then were awakened after REM o r N R E M sleep to provide
free associations and dream reports (S hevrin & Fisher, 1967). N o differences
in the dream reports were found, b u t as predicted, there were m ore “ concep
tu a l” associations after N R E M awakenings and m ore u n lik e ly or “ unusual”
associations after R EM awakenings. Flowever, the same differences were
found in o ther laboratory studies w ith o u t th e use o f any presleep s tim u li
(Fiss, E llm an, & K le in , 1969; Fiss, K le in , & B okert, 1966). In any event,
a study o f free associations after awakenings can provide no evidence about
w hether presleep s tim u li influence the process o f dream ing. M oreover, as
S h evrin and Eiser (2000) acknowledged, “ th is study d id n o t establish the
disguising fu n c tio n o f the prim ary process” (p. 1006).
Future attem pts to use sublim in al stim u la tio n to study the process o f
dream ing are n o t lik e ly to be o f any use. A w ide range o f carefully co n tro lle d
experim ents has led m ost research psychologists to conclude th a t su b lim in a l
s tim u li are lim ite d to sm all p rim in g effects fo r one o r tw o words and have
no a b ility to influence concepts; they therefore doubt th a t the processing
o f these s tim u li could have a nything to do w ith Freudian ideas (e.g., Fudin,
1999; G reenwald, 1992; G reenw ald, D raine, &. Abram s, 1996).

Freud (1900) did n o t discuss sym bolism to any great extent in the first
e d itio n o f The Interpretation o f Dreams, and as already noted, he was careful
to keep figurative th in k in g separate from the dream -w ork by m a in ta in in g
th a t th e dream -w ork sometimes makes use o f th is w aking cogn itive process.
G ive n the lack o f success in dem onstrating th a t the dream -w ork is a unique
psychological process during sleep and the success o f recent w ork on w aking
figurative thought, it is d iffic u lt to sustain the d is tin c tio n Freud trie d to
make. T he d is tin c tio n also breaks dow n w ith a close lo o k at Freud’s examples
o f the dream -w ork, w h ich make frequent use o f sexual slang, jokes, w ord
etym ologies, and proverbs. T h is appeal to the products o f w aking th o u g h t
suggests th a t the dream -work, if it exists, is an instance o f figurative though t,
as co n vin cin g ly argued by States (1987) and L a ko ff (1997).
Freud’s views on affect (i.e., em otion ) in dreams are closely tie d to his
claim s about the dream -work and have been challenged by m ore recent
findings. A cco rd in g to Freud (1900), the em otions in dreams are often
inappropriate to the content: “ I dream I am in a frig h tfu l, dangerous, repulsive
situ a tio n , b u t I feel no fear or re vu lsio n at a ll; at o ther tim es, on the contrary,
I am fille d w ith h o rro r at som ething harmless, and w ith d e lig h t at som ething
c h ild is h ” (p. 299). He attributes th is m ism atch to the fact th a t the contents o f
dreams are transform ed by displacem ents and substitutions, b u t the em otions
rem ain in place “ unaltered” (p. 299). However, laboratory studies suggest
th a t the em otions in dreams are “ overw helm ingly appropriate to the dream
co n te n t” (Foulkes, 1999, p. 68; Foulkes, S u lliva n , Kerr, & Brown, 1988;
M e rritt, S tickgold, Pace-Schott, W illia m s , & Hobson, 1994).
Freud corre ctly anticipated th a t there w ould be objections to his w ish-
fu lfillm e n t theory based on anxiety dreams and punishm ent dreams. In the
case o f anxiety dreams, he claim ed th a t they sim ply show th a t the censor
has fa ile d to disguise the wishes enough to make them acceptable. In the
case o f punishm ent dreams, he said th a t the w ish came from the censoring
agency w ith in the personality. Those ideas are plausible enough w ith in the
co n te xt o f his basic assumptions, b u t u n like actual scie n tific hypotheses,
they also make it d iffic u lt, if n o t im possible, to fa lsify the theory. H ow ever,
Freud d id n o t try to explain away, o r even discuss, the re p e titive nightm ares
th a t are the hallm arks o f posttraum atic stress disorder. Such dreams present
the greatest problem fo r the w is h -fu lfillm e n t dim ension o f his theory. I f the
theory cannot accommodate re p e titive nightm ares, th e n it is n o t necessary
to become in vo lve d in an argum ent over his theory-saving claim s about
anxiety dreams and punishm ent dreams.
Freud d id n o t consider tra u m a tic dreams, an om ission th a t is somewhat
surprising because he had been impressed sh o rtly before he w rote his book
on dreams by the role th a t ch ild h o o d traumas seemed to play in creating
n e u ro tic symptoms in adults. A lth o u g h Freud soon abandoned his traum a
theory o f the neuroses, he never abandoned his interest in traum a as a key
A C R ITIQ U E OF TRADITIONAL DREAM THEORIES 139

to the hum an psyche. T w elve years after he published The Interpretations o f
Dreams, he concluded in his first m ajor w ork on culture, Totem and Taboo,
th a t th e traum a o f p a tricid e was the startin g p o in t fo r hum an society. In
th a t volum e, he e x p lic itly debated the question o f w hether in fa n tile wishes
o r real traum a is the ce n tra l issue, concluding th a t h is to ric a l traum a, n o t
tim eless psychological processes o f hum an developm ent, m ust be the startin g
p o in t (Freud, 1912, pp. 159 -1 61).
Even so, it was n o t u n til the recurrent “war neurosis” dreams suffered
by com batants in W o rld W ar I were brought to his a tte n tio n th a t Freud
focused o n the issue o f traum a in re la tio n to dreams. A t th a t tim e he
conceded the m ain p o in t, w h ich is th a t “ it is im possible to classify as w ish
fu lfillm e n ts the dreams we have been discussing w h ic h occur in traum atic
neurosis, o r the dreams during psychoanalysis w h ich b rin g to m em ory the
psychical traumas o f ch ild h o o d ” (Freud, 1920, p. 32). H e decided th a t such
dreams reveal an attem pt at m astering overw helm ing external stim u li. I t is
in terestin g th a t th is explanation parallels w h at he o rig in a lly th o u g h t about
n e u ro tic symptoms in the 1890s.
A lth o u g h his concession clearly underm ines the theory th a t a ll dreams
are w ish fu lfillm e n ts , Freud (1933) claim ed th a t the “ exception does n o t
o verturn th e ru le ” (p. 29). H e argued th a t exceptions do n o t co n tra d ict his
theory because they draw on a deeper le vel o f the m ind, one “ beyond the
pleasure p rin c ip le ” th a t shapes m ost dreams. H is o n ly m o d ifica tio n o f his
theory was to say th a t dreams are a disguised atte m pt at w ish fu lfillm e n t;
he th e n concludes th a t “ unconscious fix a tio n to a traum a seems to be the
forem ost am ong these obstacles to the fu n c tio n o f dream ing” (Freud, p. 29)
B u t such a resolu tion o f the problem is n o t satisfactory. A s shown
in chapter 1, the accum ulation o f findings on the re p e titive dreams o f
posttraum atic stress disorder since Freud suggests th a t they cannot be so
easily isolated from dream life in general. M ore people have them , in clu d in g
victim s o f n a tu ra l disasters, tra ffic accidents, rape and assaults, and they
persist longer th a n Freud’s dismissal im plies (B arrett, 1996). T he y are experi
enced by 15% to 20% o f the wom en and m en who served in the V ietnam
W ar, and n o t on ly by veterans w ho saw d ire ct com bat. Interview s and
surveys suggest th a t young particip ants in the V ie tn a m W ar were m ore
lik e ly to suffer th is syndrome and th a t those w ho experienced the unexpected
loss o f a close frie n d were m ore vulnerable than those w ho ke p t to themselves
and form ed no close attachm ents (H artm ann, 1984). Laboratory studies
have found th a t these dreams occur in b o th R E M and Stage I I N R E M
(K ram er et al., 1987; V a n der K o lk et al., 1984).
Far fro m being an exception to a general theory, the dreams o f posttrau-
m atic stress disorder are a strong test o f the adequacy o f any new theory o f
dreams. A t th e very least, trau m atic dreams and recurren t dreams show th a t
w is h -fu lfillm e n t dreams are on ly a subset o f a ll possible dreams. These dreams

therefore stand as a re fu ta tio n o f the w is h -fu lfillm e n t theory as a general
theory. They lead to a rath er mundane idea: Just as people have some w aking
thoughts th a t are w ishful, so, too, do they have some dreams th a t are
w ishful. T his conclusion fits w e ll w ith the neurocognitive m odel o u tlin e d
in chapter 1.
The m ain defense o f the w is h -fu lfillm e n t theory has come fro m Solms
(1997). He sees his conclusion th a t the dopam inergic system is the trigger
fo r dreams as support fo r Freud’s idea, because this system is the basis fo r
“ appetitive interests.” However, his strong emphasis on the dopam inergic
system does n o t speak to posttraum atic dreams. I t is greeted w ith skepticism
by o ther neuropsychologists and neurophysiologists because dopam ine p ro
d u ctio n is about the same in w aking and R E M and is probably o n ly one
aspect o f a com plex neurochem ical m ixtu re (D o ricch i & V io la n i, 2000;
G ottesm ann, 1999, 2000; Perry & Piggott, 2000). M oreover, dopam inergic
blockers do n o t elim in a te dreams (H obson et al., 2000a, pp. 1028-1029);
in a com plex system lik e the neural netw o rk fo r dream ing, it is u n lik e ly
th a t only one lo ca tio n can stim ulate it in to action:
A s neuroim aging studies make clear, dream ing is a com plex process
occurring in a system o f m u ltip le in te ra c tin g u n its across the b ra in . In
such a d istrib u te d system, lesion studies cannot provide any means fo r
deciding on a single lo c a tio n as th e co n tro lle r, because in fa c t there
need be no such clearly defined m odule. (Bednar, 2000, p. 908)
Freud (1900, pp. 3 3 7 -3 3 8 ) though t th a t people forget m ost o f th e ir

dreams due to a h yp o th e tica l cogn itive process called “ repression,” a process
fo r w h ich little or no co nvincing experim ental evidence exists (Loftus,
Joslyn, & Polage, 1998; Loftus & Ketcham , 1994; Loftus & Polage, 1999).
M oreover, investigations o f the relatio nship between frequency o f re ca ll
and various personality and co g n itive variables cast doubt on the n o tio n
th a t any process o f den ial or self-censorship is in volve d in dream fo rg e ttin g
(C ohen, 1979; Goodenough, 1991). S im ila rly, the results from several d iffe r
e n t laboratory studies dem onstrated th a t the classic m emory variables—
recency, length, and in te n sity— best p redict w h ich dreams reported after
awakenings in the n ig h t also are recalled the next m orning (Baekland &
Lasky, 1968; M eier et al., 1968; Strauch, 1969; T rin d e r & Kram er, 1971).
Freud’s (1900) m ost controversial cla im is th a t “ impressions fro m the
earliest years o f our life can appear in our dreams, w h ich do n o t seem to
be at the disposal o f our m emory w hen we are awake” (p. 144). O nce again,
the memories can be varied in nature, b u t his greatest emphasis is on in fa n tile
sexual desires from “ up to about the end o f our th ird year” ; he dem onstrated
his p o in t by cla im in g th a t the “ nakedness-dreams” experienced by m any
adults are “exhibition-dream s,” w h ich are based on ch ild ish desires to prance
around naked (Freud, 1900, pp. 188 -1 89). Such an assertion about early
memories is h ig h ly u n lik e ly in the lig h t o f m odern-day research o n m emory,
w h ic h shows th a t few o r no conscious episodic recollections occur from
before age 3 (H ow e, 2000; Loftus &. Ketcham , 1994).
In w h at may be his m ost sweeping and elegant construction, w h ich
builds on the w is h -fu lfillm e n t theory and the fa ct th a t m ost dreams are
forg otten, Freud theorized th a t dreams are the “ guardians” o f sleep, arising
to deal w ith any b o d ily urges th a t may develop d u rin g the n ig h t (1900, p.
180). H e used “dreams o f convenience,” such as those o f going to the
bathroom or ha vin g a glass o f water, as sim ple examples o f h ow dreams
preserve sleep by p ro vid in g a h a llu cin a to ry satisfaction to an urge. H ow ever,
his greatest emphasis was on the role o f dreams in preserving sleep in the
face o f in fa n tile sexual urges.
T h e once plausible idea th a t dreams are the guardians o f sleep is now
con tra d icte d by several d iffe re n t kinds o f findings. First, the frequency and
re g u la rity o f dream ing in m ost people suggests th a t the process cannot be
p rim a rily a way to deal w ith urges th a t emerge episodically d u rin g sleep.
Second, the system atic study o f dream co n te n t thro ugh laboratory awaken
ings contradicts Freud’s theory because dream life so rarely includes the
less thre ate nin g b u t nonetheless urgent desires, such as hunger and th irs t.
A lth o u g h dream reports sometimes refer to eating and d rin k in g , the s trikin g
fa ct is how in frequ ent those incidents are if the fu n c tio n o f dreams is to
guard sleep against wishes th a t may lead to awakenings. T h ird , as discussed
in chapter 1, there is every reason to believe th a t preschool c h ild re n seldom
dream, b u t they sleep soundly nonetheless (Foulkes, 1982, 1999). Fourth,
leucotom ized schizophrenics show norm al sleep in the laboratory, b u t they
rarely re p o rt dreams even from REM awakenings (Jus et al., 1973).
In the face o f these objections, the m ain defense o f the guardian-of-
sleep theory has come from a claim by Solms (1997) th a t the neural netw o rk
may in vo lve the “ backward p ro je ctio n ” o f impulses arising in the dopam iner
gic system to the in fe rio r parie ta l lobes and visual association cortex, thereby
preserving sleep. There is, however, little o r no evidence th a t such a mecha
nism is responsible fo r dream ing (A ntro bus, 2000a, p. 905; D o ric c h i &
V io la n i, 2000). I t is m ore like ly, as stated earlier in th is section, th a t no
single area generates a ll dreams (Bednar, 2000).
Solms (1997, p. 165) also defended Freud’s fu n ctio n a l hypothesis w ith
his fin d in g th a t study participants who reported the cessation o f dream ing
m ore o fte n said th a t they had disrupted sleep th a n the c o n tro l sample did.
H ow ever, the findings are n o t impressive in th a t 51% o f the 101 particip ants
w ith global loss o f dream ing indicated th a t th e ir sleep was n o t disrupted.
I f dream ing is necessary to preserve sleep, th e n v irtu a lly everyone reporting
global cessation o f dream ing should be suffering far more from disturbed
sleep th a n Solm s’s results suggest they do.

Because standard em pirical m ethods have n o t been able to show support
fo r any aspect o f Freud’s theory, his conclusions about the nature o f dreams
rest exclusively on his m ethod o f free association, in w h ich the dreamer
produces u n c ritic a l, unreflective trains o f though t to each part o f the dream.
Freud assumed th a t free associations reveal the la te n t wishes on w h ich
dreams are based. T he seeming discovery o f la te n t dream -thoughts through
free association th e n led to his inference th a t the co g n itive processes called
the dream -w ork transform these la te n t wishes in to the m anifest dream
content.
However, there is no evidence th a t free association has any specific
usefulness fo r exam ining Freud’s hypotheses, even though it seems to be
h e lp fu l in getting people to ta lk about th e ir em otional memories and current
concerns. A s Fisher and Greenberg (1977) noted in th e ir first assessment
o f Freud’s w ork on dreams, “ there is n o t a shred o f em pirical o r reliable
evidence th a t they [free associations] provide a unique ‘true’ so lu tio n con
cerning w hat is contained in the dream” (p. 66). In a dd ition , it can be
recalled fro m chapter 2 th a t the large-scale attem pt by Foulkes (1978) to
make use o f free associations to understand dreams collected in the laboratory
setting ended w ith the conclusion th a t the m ethod is in h e re n tly a rbitrary
(Foulkes, 1996a, p. 617).
C ontrary to the cla im th a t the m ethod is free o f any suggestive influence
by the psychotherapist, experim ental evidence indicates th a t subtle sugges
tions from an e xp e rim e nte r-th era pist can falsely convince many people on
the basis o f dream interpretations th a t they were once lo st or abandoned
as young ch ild re n (M azzoni & Loftus, 1998; M azzoni et al., 1999). These
and many other findings on the power o f suggestion in a therapeutic co n te xt
(Ofshe & W atters, 1994) take on greater im portance w hen Freud’s (1900,
pp. 114-119) several reports o f arguments w ith patients concerning the
w ishful and in fa n tile bases o f th e ir dreams are added to the picture. W h a t
Freud saw as overcom ing “ resistance” can be understood from the vantage
p o in t o f social psychology as a process o f persuasion and conversion. T h is
fin d in g does n o t mean th a t a ll psychoanalytic sessions have been shown to
be exercises in suggestion. H ow ever, it does mean th a t the burden o f p ro o f
' is now on Freudians to dem onstrate th a t any therapeutic data they use to
make claim s about dreams are n o t confounded by th is extrem ely im p o r
ta n t variable.
Despite the fa ilu re o f a ll o f his specific hypotheses, Freud deserves
considerable cre d it fo r cham pioning the general idea developed in the 19th
century th a t dreams have personal psychological m eaning. I t also seems
lik e ly th a t his idea o f w ish fu lfillm e n t holds true fo r some unknow n num ber
o f dreams. In ad d itio n , he is responsible fo r the idea th a t dreams may be
the product o f figurative thought.
143
J U N G ’S T H E O R Y O F D R E A M S
T he Jungian theory o f dreams is based on fo u r m ain ideas, w h ich can

be addressed w ith m odern-day research on m etaphor, on the one hand, and
the large lite ra tu re on dream content, on the other. Jung’s (1963, 1974)
best-know n and m ost ce n tra l idea is th a t h ig h ly sig n ifica n t dreams are the
products o f a collective unconscious, w h ich contains the in h e rite d e xp e rie n tia l
record o f the hum an species in the form o f archetypes, or h ig h ly energized
patterns o r concepts th a t m ust be expressed through the personality. H o w
ever, as a com prehensive analysis and synthesis by N e her (1996) argued,
th is concept is u n scie n tific because it (a) is based o n the discredited n o tio n
o f the inheritance o f acquired characteristics; (b) does n o t allo w fo r the
v a ria tio n in specific archetypes th a t w ould be expected on the basis o f
m odern-day genetics; and (c) is n o t grounded in a co n vin cin g e lim in a tio n
o f the possible influence o f socialization and culture in the personal, thera
peutic, and cross-cultural anecdotes on w h ich it is based.
In ad d itio n , as the firs t p o in t in N eher’s (1996) c ritiq u e im plies, the
concept is n o t able to escape the charge o f c irc u la rity because the origins
o f the c o lle ctive unconscious are said to be in repeated hum an experience,
w h ic h is the phenom enon the concept is supposed to explain. I f experience
is the basis fo r the co lle c tiv e unconscious, there is no need to invoke a
co lle ctive unconscious to explain com m onalities o f experience th a t are m ost
lik e ly based o n the sim ila r hum an situations th a t recur in each in d iv id u a l
life tim e . T h is vie w p o in t also can better explain the variatio ns in th in k in g
styles th a t are found fro m culture to culture (R ogoff, 1990).
Second, Jung argued th a t the archetypes o f the co lle ctive unconscious
express themselves thro u g h a set o f in h e rite d symbols th a t also appear in
m yths, religious cerem onies, and o th e r w aking practices. Thus, the m ain
focus o f Jungian dream analysts is on in te rp re tin g these symbols using b o th
in d iv id u a l dreams and c u ltu ra l parallels. Jung’s observation o f some com m on
a lity in dream co n te n t across in d ivid u a ls and cultures is m ore parsim oniously
and plausibly encompassed by the idea th a t m etaphorical concepts are
acquired through b o th developm ental experiences shared by a ll hum an
beings and gradual lin g u is tic socialization in to the huge treasure trove o f
conceptual m etaphors th a t are p a rt o f a group’s c u ltu ra l heritage (G ibbs,
1994; Lakoff, 1987; L a k o ff & T urner, 1989). T h is idea is supported by
findings show ing sim ila ritie s in conceptual m etaphors in m any d iffe re n t
cultures, in clu d in g C h in a and Japan (L a ko ff &. Johnson, 1999; N om ura,
1996; Yu, 1999).
T he usefulness o f th in k in g about sym bolic in terpretations as m eta
p h o ric analyses can be seen by lo o kin g at the dream th a t Jung (1963) claim s
to be the basis fo r his idea o f a co lle ctive unconscious. In th a t dream, from
a tim e w hen he was ha vin g grave doubts about Freudian theory, Jung found

h im se lf exploring a house th a t belonged to h im b u t did n o t seem at a ll
fa m ilia r. H e was impressed by the elegant decor o f the upper floor, and th e n
descended to a low er floor w ith m edieval furnishings. H e n e xt saw a heavy
door th a t led to a stone stairway and an ancient vaulted cellar. O n the
ce lla r flo o r he saw a stone slab leading dow n another set o f stairs to “ a lo w
cave cut in to the rock,” w h ich contained rem ains from a p rim itiv e culture
(Jung, 1963, p. 159). Jung interpreted the dream as “ a structural diagram
o f th e hum an psyche,” w ith the deepest le ve l o f the house representing a
co lle ctive unconscious o f an im personal nature (Jung, 1963, p. 161).
T his in te rp re ta tio n makes “ in tu itiv e ” sense to many people because it
is based on a shared conceptual m etaphor: T he “ m in d is a container” (Lakoff,
1987; L a ko ff & Johnson, 1999). T he same conceptual m etaphor makes it
possible to say th a t there are “ skeletons in the closet” or “ bodies buried in
the basement” w hen a person has a secret, o r to say “ there is no one hom e”
w hen a person is absentm inded or n o t in te ra ctin g properly, o r to characterize
m ental illness by saying a person has “ bats in his b e lfry” or “bees in her
bon net.” In fact, a ll o f the “fu n c tio n a l” or “ subjective” symbols said by
Jungians to represent the psyche, or aspects o f it, can be lin ke d to one or
m ore o f several conceptual metaphors.
Jung’s th ird m ajor idea is th a t m ost dreams, especially those w ith roots
in th e co lle ctive unconscious, have a compensatory fu n c tio n — th a t is, they
express those aspects o f the personality, in clu d in g the archetypes, th a t are
n o t adequately developed in w aking life . T h is idea is d iffic u lt to support or
refute in a d e fin itive way because there may be subtle form s o f com pensation,
even in dreams th a t do n o t seem com pensatory on the basis o f objective
m ethods. S till, the idea seems to be contradicted by every relevant systematic
study since the beginning o f m odern-day dream research in the late 19th
century, when psychologists who w rote dow n th e ir ow n dreams found consid
erable c o n tin u ity between dream content and w aking co g n itio n (C alkins,
1893; W eed & H allam , 1896). T h is fin d in g was repeated in nonlaboratory
investigations summarized by Fisher and G reenberg (1977, 1996), and it
finds fu rth e r support in laboratory studies by Fiss (1983, 1986) th a t used
c lin ic a l assessments as w e ll as in tw o laboratory studies th a t analyzed correla
tio n s between dream content and objective personality measures (Foulkes
et al., 1969; Foulkes & Rechtschaffen, 1964). In b o th o f the laboratory
corre latio nal studies, the young m en w ho had the m ost unpleasant dreams
also tended to have the highest scores on M M P I psychopathology indicators.
A s stressed throughout th is book, the c o n tin u ity between dream con
te n t and w aking life is one o f the m ost s trik in g findings from content analysis
studies by C a lv in S. H a ll and his co-workers (H a ll & N ordby, 1972). People
dream most often about the people and interests th a t preoccupy them in
w aking life . They show the m ost aggression in dreams tow ard the people
w ith whom they have the m ost c o n flic t in w aking life . T he results are so
A CRITIQUE OF TRADITIONAL DREAM THEORIES 145

consistent fo r these kinds o f c o n tin u itie s th a t H a ll adopted the term continu-
ity principle to contrast his findings w ith Jung’s com pensation hypothesis.
H a ll’s b lin d analyses o f the dreams o f a c h ild m olester, a n e u ro tic p a tie n t
in psychotherapy, Franz Kafka, and numerous average people w ho kept
dream journals provide strong evidence fo r th is alte rn a tive hypothesis (B e ll
&. H a ll, 1971; D om hoff, 1996; H a ll &. L in d , 1970). T h is evidence is now
supplem ented by the new findings presented in chapter 5.
F ourth, and fin a lly , Jung claim ed th a t gradual changes in dream con te n t
occur beginning in the m iddle years o f adu lt life th a t reflect the psychological
need fo r the “ in d iv id u a tio n ” and “ in te g ra tio n ” o f the personality, under the
d ire c tio n o f the “ s e lf’ archetype. H ow ever, as m entioned in chapter 1, there
is considerable evidence th a t adults, u n like ch ild re n , are consistent in w hat
they dream about over m onths, years, or decades. T he evidence is o f tw o
types: cross-sectional and lo n g itu d in a l. Several cross-sectional studies, m ost
in the U n ite d States, but one in Canada and one in S w itzerland, dem on
strated th a t dream c o n te n t shows consistency, n o t change, as people grow
older, w ith the possible exception o f declines in aggression and negative
em otions (Brenneis, 1975; C ote et al., 1996; D om hoff, 1996; H a ll &.
D om hoff, 1963b, 1964; H ow e &. B lic k , 1983; Inge Strauch, D epartm ent o f
Psychology, U n iv e rs ity o f Zurich, personal com m unication, A p ril 5, 2000;
Zepelin, 1980, 1981).
A lth o u g h the lo n g itu d in a l studies were n o t large in num ber, they were
sim ila r in th e ir results fo r b o th ongoing dream journals (D om hoff, 1996;
D o m h o ff &. Schneider, 1998; H a ll <&. N ordby, 1972; S m ith & H a ll, 1964)
and 2-week journals collected 10 to 17 years apart from 21 w om en (L o rtie -
Lussier et al., 2000). T he analyses o f the Barb Sanders series in chapter 5
now add an im p o rta n t new dim ension to th is body o f evidence because o f
the num ber o f dreams th a t were studied and the d iffe re n t types o f analyses
th a t were conducted. Few changes occurred in m ost o f the coding categories
o r in various subseries, even though the dreams cover the years w hen the
process o f in d iv id u a tio n is supposed to be unfoldin g.
W ith in th is co n te xt o f consistency, a study o f dreams from wom en
before, during, and after menopause is o f special in terest because it was
designed to test ideas derived from Jungian theory concerning changes in
dream co n te n t during and after menopause (A b e l, 1994). T he investigator
created seven th e o re tica l scales derived from Jungian theory th a t were used
by three coders naive about Jungian theory and b lin d as to the age o f the
dreamer. Even m ore im p o rta n t, an expert on Jungian theory coded a ll dreams
o n a scale fo r the degree to w h ich the dreams expressed archetypal symbols.
T h irty -tw o o f the w om en were s till m enstruating (average age = 39.6);
24 w om en were perim enopausal (average age = 49.2); and 20 w om en were
postm enopausal (average age = 58.6). W om en in the firs t group averaged
32 dream reports over a 1-m o n th period; w om en in the second group averaged

21 dream reports; and women in the th ird group averaged 16 reports. T h is
decline in dream recall parallels the decrease in dream recall w ith age in a
study o f 2,328 adults ages 17 to 92 (G iam bra, Jung, &. Grodsky, 1996). T he
o n ly sig nifica nt difference in dream co n te n t concerned an “ in itia tio n ” scale,
b u t the difference was n o t fo r the predicted group. C o ntrary to expectations,
the codings by the Jungian expert d id n o t show any difference in the degree
o f sym bolic expression in the three samples.
O nce again, then, the findings relevant to the theory under investiga-
tio n provide no support fo r any o f its m ost im p o rta n t claim s. In the end,
Jungian dream theory boils dow n to the idea th a t dreams can be understood
in term s o f w aking conceptual metaphors. N o evidence supports th a t claim ,
and no rules or guidelines determ ine w h ich conceptual m etaphors should
be applied to w h ich dreams. As the study o f metamorphoses in the Sanders
series shows, developing such rules and evidence w ill n o t be easy.
A C T IV A T IO N -S Y N T H E S IS TH E O R Y
A s the firs t m odern-day neuropsychological theory o f dreams th a t draws

on sleep research, activatio n-synth esis theory is a forerunner o f the k in d
o f neurocognitive m odel presented in th is book. How ever, it differs from a
neurocognitive theory in th a t it starts w ith studies o f b ra in stem lesions in
cats and extrapolates to hum an dream ing w ith o u t considering any o f the
findings concerning dream ing and dream co n te n t th a t are based on labora
to ry awakenings (Foulkes, 1966, 1985). T he a ctivatio n-synth esis theory is
distinguished by its emphasis on cells in one region in the pons as th e on ly
trigger fo r dream ing and as the m ain determ inant o f some o f the unusual
form al aspects o f dreams. A cco rd in g to the o rig in a l version o f th is theory,
w h ich became popular as the antithesis o f Freudian theory, a dream is a
catch-as-catch-can synthesis by the forebrain, w h ich is “ m aking the best o f
a bad job in producing even p a rtia lly coherent dream imagery from the
re la tive ly noisy signals sent up to it fro m the brainstem ” (H obson & M c-
C arley, 1977, p. 1347). T he theory offers no suggestion th a t any co g n itive
processes are operative. M ore specifically, phasic stim u la tio n from th e b ra in
stem, p rim a rily in the form o f pon to -g e n icu lo -o ccip ito (P G O ) waves, is
responsible fo r sudden scene changes, unusual juxtapositions, and other
im probable constructions in dreams (Hobson, 1988; M am elak & Hobson,
1989). T he a c tiv a tio n o f the visual system in REM sleep is assumed to be
“ fo rm a lly sim ila r to th a t o f the w aking state,” w h ich is said to account fo r
“ the c la rity o f our dream visio n ” (Hobson, 1988, p. 205).
Even some types o f dream co n te n t are claim ed to be d ire ct reactions
to b ra in stem signals, such as being unable to move w h ile being chased,
w h ich is described as an accurate co rtic a l reading o f the contradictory state

created by the h ig h a c tiv a tio n o f the m o to r-p a tte rn generator in c o n ju n c tio n
w ith the paralyzed state o f spinal neurons. S im ila rly , “ flyin g dreams may
thus be a logical, direct, and unsym bolic way o f synthesizing in fo rm a tio n
generated endogenously by the vestibular system in D [dream ing] sleep”
(H obson & M cG arley, 1977, p. 1339). T he poor re ca ll o f dreams is said to
reflect a “ state-dependent amnesia,” w h ich may be a result o f the low
le ve l o f “ am inergic neuronal a c tiv ity and the resulting effects on second
messengers and m acrom olecules” (H obson & M cC arley, 1977, p. 1347).
From the p o in t o f view o f m any o f the early laboratory dream research
ers, th is theory was contradicted by several types o f data th a t already had
been collected in laboratories in the 20 years before it appeared. First, the
presence o f dream ing in m any awakenings fro m N R E M sleep is strong
evidence against an exclusive emphasis on R E M sleep as the co n te xt fo r
dream ing (Foulkes, 1962; Foulkes & S chm idt, 1983; Foulkes & V ogel, 1965;
H erm an, E llm an, & Roffwarg, 1978). A lth o u g h these studies show th a t
m any N R E M dream reports are less “ dream like” and more “ th o u g h tlik e ”
th a n ty p ic a l R EM dream reports, more th a n enough N R E M reports co n ta in
fu ll-b lo w n dream co n te n t to co n tra d ict the s tric t equation o f R E M sleep
and dream ing. Indeed, th is evidence was so co n vin cin g to m ost dream
researchers by the late 1960s th a t they had already abandoned the R E M -
equals-dream ing equation th a t was central to the o rig in a l version o f a ctiva
tio n -syn th e sis theory (Berger, 1967, 1969; Foulkes, 1966; H a ll, 1967).
Second, dreams at sleep onset, long before there are signs o f REM ,
co n tra d ict the theory (V ogel, 1991). A lth o u g h dream like m ental a c tiv ity
is b rie f during the tra n s itio n to sleep, it is extrem ely im p o rta n t th e o re tica lly
because th e h ig h ly regular sequence o f m ental changes discovered in the
laboratory does n o t correlate strongly w ith the physiological changes th a t
index the tra n s itio n fro m w aking to sleep. For exam ple, h a llu cin a to ry im ag
ery can even occur w hen the EEG p a tte rn s till indicates wakefulness
(Foulkes, 1985, pp. 7 0 -7 1 ).
These longstanding results were la te r supplem ented by a c tiv a tio n -
synthesis theorists themselves in a study o f 16 particip ants over a 2-week
period, w ho were paged w h ile awake o r aroused by means o f the N ig h tca p
w h ile sleeping in th e ir homes. T he study found th a t the prevalence o f
“h a llu c in a to ry co n te n t” ranged from 3.6% in quie t w aking to the higher
figures o f 35% at sleep onset, 60% in N R E M , and 82% in REM (Fosse et
al., 2001, p. 33). T h e percentages fo r the N R E M and R E M awakenings are
sim ila r to w hat has been found in laboratory studies, b u t are o n ly about
h a lf o f w h a t has been found fo r sleep onset thro u g h carefully p in p o in te d
laboratory awakenings (Foulkes, 1985; Foulkes & V ogel, 1965; Vogel, 1991;
V ogel, Barrow clough, & G iesler, 1972).
A th ird type o f evidence against a ctiva tio n -syn th e sis theory comes
fro m studies th a t attem pted to lin k a range o f phasic a ctivitie s during R E M —

such as bursts o f eye m ovem ents, muscle a c tiv ity , or bursts o f theta waves—
to dream co n te n t through im m ediate awakenings w hen the signs appeared.
T he findings o f some o f the studies suggest th a t dream ing is m ore v iv id
during phasic events, thereby supporting the idea th a t the level o f a c tiv a tio n
may be indexed by them . H ow ever, there is little or no conn ection to dream
features or dream co n te n t according to m ost reviewers (A ntrobus, 2000b;
Foulkes, 1985; P ivik, 1986, 2000). S till, the relatio nship is described as
“weak b u t consistently p o sitive ” by activatio n-synth esis theorists (H obson
et al., 2000b, p. 799).
Fourth, the neuroim aging and lesion studies discussed in chapter 1
co n tra d ict the passive role assigned to the forebrain in the first version o f
activatio n-synth esis theory. They show th a t the forebrain is m ore active
th a n the theory assumes and th a t it is selectively active in ways n o t a n tic i-
pated by the theory. In ad d itio n , and contrary to activatio n-synth esis theory,
the prim ary visual cortex and m oto r cortex do n o t seem to be part o f the
neural substrate fo r dream ing, according to the lesion studies, and they are
n o t active during REM sleep, according to the neuroim aging studies (C o n
d u it, C rew ther, & Colem an, 2000, p. 925).
In the face o f these accum ulated findings, along w ith strong new
evidence fo r the presence o f full-fledged dream ing in N R E M ju st before
m orning awakenings (A ntro bus e t al., 1995; C icogna et al., 1998),
activatio n-synth esis theory has been altered and expanded in several ways.
In p articular, the forebrain has been given a m uch larger role in regulating
the pon tin e a ctiva tio n responsible fo r REM and in shaping dream content,
especially through lim b ic structures (H obson e t al., 1998, 2000b). T he
revised theory also includes greater a tte n tio n to tw o factors th a t were o n ly
b rie fly noted in the o rig in a l fo rm u la tio n : the “ in p u t source” and the nature
o f the neurochem ical “ m od u la tio n .”
T he expanded m odel has three dim ensions th a t are m eant to account
fo r a ll states o f consciousness, n o t ju st dream ing: activation, input, and
modulation (A IM ). A c tiv a tio n now refers to b o th to ta l and regional b ra in -
a ctiva tio n levels. In p u t concerns the degree to w h ich a ctiva tio n is being
generated in te rn a lly or externally. M o d u la tio n refers to th e ra tio o f am inergic
(i.e., serotonin and norep inep hrine) to cholin erg ic (i.e., acetylcholine) neu
rom odulators in each consciousness state. T he ra tio is h ig h in w aking, but
it is reversed during REM , w hen th e levels o f serotonin and norepinephrine
fa ll to near zero (Fosse et al., 2001, p. 30; Hobson et al., 2000b, p. 805).
In the fu ll A IM m odel, w aking is said to be characterized by h ig h levels o f
b ra in a ctiva tio n , external sources o f in p u t, and am inergic neurom odulation,
whereas R EM is characterized by h ig h levels o f b ra in a ctiva tio n , in te rn a l
sources o f in put, and ch o lin e rg ic m odulation. N R E M is characterized by
lo w levels o f brain a ctiva tio n , in te rn a l sources o f in p u t, and a m ixtu re o f
am inergic and cholin erg ic m odulation. The m odel can account fo r a va rie ty
o f unusual states o f consciousness, such as R E M sleep-behavior disorder,
sleep paralysis upon awakening, and drug-induced h a llu cin a tio n s.
C h o lin e rg ic neurom odulation during R EM , thro ugh its s tim u la tio n o f
PG O a c tiv ity , is assumed to be the prim ary causal fa cto r in accounting fo r
the “ h a llu cin a to ry” im agery o f dream ing. T he “ u b iq u ity o f m o tio n ” in dreams
is said to be caused by the h ig h le vel o f a c tiv ity in basal ganglia during
R E M , and p lo t d is c o n tin u ity and in co n g ru ity are the result o f “ deficient
executive functions in clu d in g w orking m em ory” (H obson et al., 2000a, p.
1030). D ream ing in N R E M is now a ttrib u te d to subtle s tim u la tio n by the
same regions o f the b ra in stem th a t generate R EM . T h is s tim u la tio n is
sometimes called “ covert R EM sleep” (H obson et al., 2000b; N ielsen, 2000a,
2000b). In o ther words, the revised theory has som ewhat less emphasis on
the R E M stage, b u t it has a c o n tin u in g emphasis on b ra in stem stim u la tio n
from the p o n tin e tegm entum as the key to dream ing. T o make th is p o in t,
H obson (2000) argued th a t “ a ll sleep is R E M sleep (m ore or less),” a state
m ent th a t im plies th a t states o f sleep are n o t as discrete as they were once
th o u g h t to be (p. 952). T he fa ct th a t re la tiv e ly few dreams are remembered
is now explained as an “ organic amnesia” caused by the com bined effects
o f low am inergic levels and decreased a c tiv ity in the dorsolateral prefron
ta l cortex.
These alterations narrow the distance between a ctiva tio n -syn th e sis
theory and a neurocognitive m odel on issues concerning the neural substrate
fo r dream ing. How ever, those w ho have collected dreams from N R E M
have noted th a t no d ire ct evidence connects N R E M dream co n te n t to any
p o te n tia l indicators o f co ve rt R E M sleep (B o sin e lli & C icogna, 2000; C ip o lli,
2000; Feinberg, 2000; V ogel, 2000, p. 1015). These researchers especially
believe th a t the large num ber o f dream reports fro m spontaneous N R E M
m orn ing awakenings (C icogna et al., 1998) are probably best explained by
general b ra in a ctiva tio n . M oreover, a research group th a t has collected
dreams a t sleep onset has suggested th a t the rem nants o f w aking a ctiva tio n
are a m ore lik e ly e xp lana tion fo r these dreams th a n any s tim u la tio n from
covert R E M sleep (O g ilv ie , T akeuchi, & M urphy, 2000). T here are in d ica
tions th a t activa tio n -syn th e sis theorists concur w ith a prim ary emphasis on
a c tiv a tio n fo r sleep onset and spontaneous N R E M m orn ing awakenings, as
in the com m ent th a t th e ir m odel works best fo r “ intense dream ing” in the
“ first 4 to 6 hours o f the n ig h t” (H obson, 2000, p. 951; H obson et al., 2000a,
p. 1023).
T h e gap could narrow even fu rth e r if neurophysiological critics o f
a ctiva tio n -syn th e sis theory are rig h t th a t the theory s till overstates the role
o f the p o n tin e tegm entum in R E M regulation by n o t g ivin g enough w eight
to grow ing evidence fo r c o n tro l o f th is sleep state by the hypothalam us
(M o rriso n & Sanford, 2000; Salin-Pascual, G erashchenko, & S hirom ani,
2000). R egulation by the hypothalam us includes c o n tro l o f the PG O waves

th a t are central in a ctivatio n-synth esis theory. A s Steriade (2000) argued,
hypothalam ic and forebrain structures “ may be the m ost effective in d riv in g
PG O neuronal generators” (p. 108). T he gap could alm ost be closed in
term s o f differences o f o p in io n on the nature o f the neural substrate fo r
dream ing if other neurophysiologists are correct th a t the theory has too
exclusive an emphasis on cholin erg ic systems in the creation o f dream
m entation. A s G ottesm ann (1999, 2000), Perry, W alker, Grace, and Perry
(1999), and Perry and P iggott (2000) argued on the basis o f a range o f
evidence in anim als and humans, it seems more lik e ly th a t, at the least, it
is the com bination o f acetylcholine and dopam ine, in the absence o f seroto
n in and norepinephrine, th a t modulates REM . A s the m ixtu re becomes
m ore com plex, neurom odulation shades tow ard the emphasis on “ general
a ctiv a tio n ” favored by co g n itive ly oriented theorists.
O ver and beyond these s tric tly neurophysiological issues, the A IM
m odel has four o ther problem s, a ll o f w h ich relate to the c o n tin u in g neglect
o f the cognitive dim ension, w h ich rem ains underdeveloped in the theory.
T he first o f these a d d itio n a l problem s concerns the re la tive absence o f
dream ing in young ch ild re n , w h ich is denied by a ctivatio n-synth esis th e o
rists. T o deal w ith young c h ild re n ’s lo w re ca ll from REM periods, the
researchers could have m odified th e ir view to say th a t R E M is necessary
b u t n o t sufficient, as they d id on the basis o f the findings from adults w ith
b ra in lesions. Instead o f acknow ledging th a t cogn itive developm ent is also
necessary fo r dream ing, they argued th a t the ch ild re n in Foulkes’s (1982;
see also Foulkes, H o llifie ld , S u lliva n , Bradley, & Terry, 1990) studies fe lt
uncom fortable and in h ib ite d in the laboratory (H obson et al., 2000b). T he y
presented th e ir ow n evidence o f full-fledged dreams from preschool ch ild re n
on the basis o f hom e-reported dreams collected by parents (Resnick, S tick-
gold, Rittenhouse, & Hobson, 1994). Some o f the dreams were collected after
having the ch ild re n te ll themselves at bedtim e th a t they w ould remember a
dream in the m orning, w h ich Foulkes (1996b, 1999) sees as an im p lic it
pressure to com ply th a t may have led to made-up dreams.
As if to underscore th e ir differences w ith Foulkes’s emphasis on the
need fo r a certain le vel o f co g n itive developm ent fo r dream ing to occur,
the activatio n-synth esis theorists assert th a t they can im agine dream ing
in neonates:
S im ila rly, we specifically suggest th a t the hum an neonate, spending as

it does m ore th a n 50 percent o f its tim e in R EM sleep, is havin g
indescribable b u t nevertheless real o n e iric experiences. A n in fa n t’s w ak
ing experience rem ains essentially indescribable and speculative to us
older persons b u t we do n o t doubt th a t infants enjoy some sort o f w aking
conscious experience. For us, it is n o t at a ll d iffic u lt to im agine th a t
an in fa n t m ig h t be experiencing hallucinosis, em otions, and fic tiv e
kin esthetic sensations d u rin g R EM sleep. (H obson et al., 2000b, p. 803)
A CRITIQUE OF TRADITIONAL D REAM THEORIES 151

C o ntrary to these claim s, Foulkes (1982, 1999) presented detailed
evidence th a t his extensive efforts to make the ch ild re n com fortable in the
laboratory setting d id prove successful. In ad d itio n , during the second and
fo u rth years o f his study, he tested fo r the possible effects o f awakenings by
a llow ing the c h ild re n to sleep throughout the n ig h t in the laboratory. T hey
th e n reported any dreams they recalled in the m orning. T h is procedure
found n o differences w ith dreams collected after m orn ing awakenings at
hom e. Foulkes’s fin d in g means th a t any differences found in o th e r studies
between dreams collected from awakenings in the laboratory and at hom e
are due to selective recall fo r atypical dreams at hom e, rather th a n to any
alleged in h ib ito ry effect in the laboratory (Foulkes, 1979, 1982, 1996b).
T he fin d in g th a t tw o boys ages 11 to 13 w ho had low visuospatial skills
unexpectedly showed lo w levels o f REM recall also suggests th a t the issue
is la ck o f dream ing, n o t la ck o f re porting skills (Foulkes, 1982, pp. 1 8 0 -
181, 2 2 5 -2 2 6 ).
Second, the a ctiva tio n -syn th e sis theory o f dream fo rg e ttin g does n o t
consider the co g n itive factors th a t seem to be in vo lve d in lo w levels o f
dream recall. A s tric tly neurophysiological explanation does n o t take in to
account th a t w aking re ca ll also can be poor in some circum stances, such
as w hen a person’s m in d is d riftin g w h ile perform ing a ro u tin e task. From
a co g n itive p o in t o f view , the problem may be the lack o f “ an external
narrative to w h ich m em ories fo r in te rn a l events can be tie d ” (C hapm an &
U nderw ood, 2000, p. 917). M oreover, m ore fo rg e ttin g in w aking life occurs
th a n is com m only assumed (C hapm an &. U nderw ood, 2000). T h e fa ct th a t
h ig h recallers tend to have a stronger interest in dreams th a n nonrecallers
(Tonay, 1993) and th a t re ca ll can be im proved w ith tra in in g and encourage
m ent (S chredl, 2000) also suggest th a t there is a co g n itive dim ension to
dream recall.
T h ird , there are unresolved questions related to claim s by a c tiv a tio n -
synthesis theorists about “ bizarreness” in dreams. A lth o u g h a c tiv a tio n -
synthesis theorists are h ig h ly c ritic a l o f a ll c lin ic a l theories o f dreams, they
contin ue to ta lk about the norm al process o f dream ing as though it were a
psychiatric phenom enon (H artm ann , 2000). For example, they c a ll dream
im agery “ h a llu c in a to ry ” and lik e n the dream experience to “ religious conver
sion, near-death experience, fu n c tio n a l psychosis, d e liriu m , drug-induced
conditions, and o th e r altered states o f consciousness” (H obson et al., 2000b,
p. 802). T he y a ttrib u te th is alleged bizarreness to forebrain stim u la tio n from
the p o n tin e tegm entum , b u t as A n trobus (2000a) stressed, “ assumptions
about how the pons determ ines the features o f dream ing are com pletely
w ith o u t em pirical support” (p. 905). He also pointed out th a t he and his
co-workers have collected instances o f unusual dream imagery w hen phasic
a c tiv ity o rig in a tin g in the p o n tin e tegm entum was m in im a l (A n tro b u s et
al., 1995).

Several d iffe re n t scales have been developed to assess bizarre elements,
b u t there is no general agreement as to th e ir usefulness or th e ir m ethodologi
cal soundness (e.g., Bonato, M o ffitt, H offm ann, & Cuddy, 1991; H a ll, 1966b,
p. 40; O g ilvie , H u n t, Saw icki, & S am ahalskyi, 1982; Revonsuo & S a lm iva lli,
1995). T he scales developed by activatio n-synth esis theorists (Hobson,
H offm an, H elfand, & Kostner, 1987; R ittenhouse et al., 1994) emphasize
sudden shifts in scenes th a t may n o t be unique to dreams if jumps in relaxed
w aking th o u g h t are used as the relevant com parison. In ad d itio n , most o f
the scales are com prom ised because they include no co n tro l fo r dream le ngth
and th e ir re lia b ility is uncertain.
T he coding categories th a t a ctivatio n-synth esis theorists have created
fo r the study o f bizarreness are no more reliable or con vin cin g th a n any
o f the other bizarreness scales. Recognizing re lia b ility problem s w ith th e ir
firs t scale (H obson et al., 1987, pp. 161 -1 62), w h ich showed on ly 55%
agreement between tw o raters w hen used by another researcher (A b e l,
1996, p. 10), Hobson and his colleagues m odified the scale fo r la te r studies.
In a study using the revised scale on 200 dream reports w ith 4,674 lines
o f te xt, it was reported th a t tw o o f three coders designated 737 lines as
co n ta in in g bizarre features, b u t no in d ic a tio n o f the num ber o f lines on
w h ich they disagreed was provided, a piece o f in fo rm a tio n th a t w ould
make it possible to determ ine the percentage o f agreement. M oreover,
on ly 456 (62% ) o f the id e n tifie d lines were agreed on by a ll three coders
(M e rritt et al., 1994, p. 53).
In ad d itio n , the boundary between features and contents is n o t always
as clear as activatio n-synth esis theorists’ coding system suggests. W h a t they
consider a bizarre feature, such as one character changing in to another or
one character being a com bin ation o f tw o people, may be “ co n te n t” th a t
is the product o f a conceptual blend (Fauconnier, 1997; G rady et al., 1999).
T h is p ossibility is dem onstrated w ith some o f th e com posite characters in
the Barb Sanders series discussed in chapter 5. In effect, H obson’s theory
and coding system ru le out any study o f figurative though t in dreams.
T he m ost frequent bizarreness th a t a ctivatio n-synth esis theorists fin d
in dreams is in the num ber o f abrupt scene changes, w h ich occurred in 34%
o f 200 dreams in one o f th e ir studies (R ittenhouse et al., 1994). By contrast,
using a scale th a t focused on unusual a ctivitie s, unusual occurrences, dis
to rte d objects, and metamorphoses, H a ll (1966b, p. 41) found th a t on ly
10% o f 815 hom e and lab dream reports had at least one bizarre elem ent,
w ith no differences between the tw o types o f reports. M oreover, there is
far more d isco n tin u ity, d rift, and in a tte n tio n in w aking though t th a n is
im p lie d by the cla im th a t changes in dream scenes or settings are in h e re n tly
bizarre (C hapm an & Underw ood, 2000). T he releva nt com parison fo r study
ing dream bizarreness is w ith reports o f w aking memories, n o t w aking re a lity
(Bednar, 2000, p. 909; Chapm an & U nderw ood, 2000, p. 917).
153
N o t a ll studies agree th a t there are frequent d isco n tin u itie s w ith in
dream reports. In a detailed study o f th is issue, Foulkes and S chm idt (1983),
divide d REM dream reports in to a series o f “ tem poral u n its,” w h ich were
defined by the appearance o f a new a c tiv ity in th e dream, such as the
sequence o f “ com ing out o f school/opening the gate/children saying goodbye
to each o th e r/w a lkin g dow n the street” (p. 267). T he y found th a t o n ly 1
in 8 tem poral tran sition s was accom panied by a d is c o n tin u ity in b o th setting
and characters. T he y argued th a t the re la tive ly sm all d isco n tin u itie s in
dreams are consistent w ith , and probably necessary for, the considerable
degree o f n arrative and th e m a tic developm ent th a t is found in m ost R EM
reports (cf. C ip o lli & P o li, 1992). I f the findings are accurate, th e n it is
lik e ly th a t the A IM m odel overemphasizes bizarreness due to its focus on
the bra in stem as a d ire ct cause o f unusual dream features.
F ourth, and fin a lly , a ctiva tio n -syn th e sis theory has little to say about
dream co n te n t due to its la ck o f a tte n tio n to the co g n itive dim ension o f
dream ing. Its m ain proponents agree th a t the re la tive “ mundaneness” o f
dream co n te n t is com patible w ith th e ir emphasis on the bizarreness o f dream
form and th a t there is at least some in fo rm a tio n and p a tte rn in in d iv id u a l
dream journals. H ow ever, th e y raise questions about the accuracy and useful-
ness o f dream reports and about the p o ssib ility o f doing scie n tific studies o f
dream co n te n t (H obson et al., 2000a, pp. 1020-1021). C onsequently, they
do n o t take seriously the need to add a con te n t dim ension to th e ir theory.
T hey therefore have little o r n o th in g to co n trib u te to the study o f dream
m eaning a t the co g n itive level.
C o ntrary to th e ir doubts, the arguments and evidence in chapter 2
show th a t the q u a lity o f dream reports can be excelle nt, and chapters 3, 4,
and 5 dem onstrate th a t it is possible to conduct scie n tific studies o f dream
co n te n t. Several H a ll-V a n de Castle categories fo r types o f a ctivitie s easily
accom m odate the a ctiva tio n -syn th e sis theorists’ focus on sensory references
and m ovem ent in dreams b e tte r than the a ctiva tio n -syn th e sis theorists’
ow n scales. T h e categories fo r em otions are also b e tte r suited th a n th e ir
scales fo r studying the questions o f interest to them , in clu d in g th e ir concern
w ith confusion, surprise, and uncertainty.
T he way in w h ic h issues concerning form and co n te n t m ig h t be in te
grated in to the new neurocognitive m odel can be shown through tw o sepa
rate studies o f the same dream series, taken from a 3 -m o n th jo u rn a l o f 233
dream reports. T he jo u rn a l was kept out o f in te lle c tu a l cu rio sity in the
summer o f 1939 by a 46-year-old na tu ra l scientist w ho had no tra in in g in
psychology or psychiatry and no investm ent in any dream theory. H e w rote
in the preface th a t he was a frequent dreamer who had been frightened by
some o f his nightm ares as a c h ild and th a t he wanted to see fo r h im se lf if there
was a nything to the general claim s by Freud. T he jo u rn a l is excep tionally
in terestin g in th a t it includes drawings th a t relate to m any o f the dreams.
For those who w ant to see the dream reports fo r themselves, they are available
under the name “T he N a tu ra l S cie n tist” at http://w w w .D ream B ank.net.
Hobson purchased the dream jo u rn a l from a m edical book catalog in
1980; he used the drawings to study the sensory references, fic tiv e m ovem ent,
and bizarre features in the dream reports, claim ing th a t his findings showed
the influence o f p o n tin e stim u la tio n on dream features (Hobson, 1988;
Hobson et al., 1987). From his th e o re tica l vantage p o in t, the dreams seemed
to be peculiar and incongruous, as indexed by sudden scene changes or
uncertainties as to the id e n tity o f people o r settings. However, no evidence
lin k e d these features to the neural ne tw o rk fo r dream ing except th e ir appar
ent frequency.
The same dream jo u rn a l yielded d iffe re n t in fo rm a tio n w hen it was
studied fro m the con te n t p o in t o f vie w by Adam Schneider, as first reported
in D o m ho ff (1996, pp. 147 -1 50). First, as shown in T able 6.1, w hen the
firs t 93 reports w ith 50 or m ore words are compared w ith the second 93,
the co n te n t is s trik in g ly consistent over ju st th is short period. Second, a
b lin d analysis o f the characters, social interactions, settings, and em otions
in the dreams provides a good p o rtra it o f the dream er’s w aking concerns
and interests, as dem onstrated thro u g h a la te r com parison o f the dream
findings w ith a four-page obitu ary th a t contains personal in fo rm a tio n .
For example, as shown in the h-profile in Figure 6.1, the dream er scores
lo w on aggressiveness and even low er on dream er-involved aggression and
T A B L E 6.1
C o n s is te n c y in th e N a tu ra l S c ie n tis t’s 3 -M o n th D re a m J o u rn a l
Total 1s t set 2nd set
(n= 187) (n = 93) (n = 93)
Characters
C haracters per dream 3.08 3.08 3.02
M ale/fem ale percent 70 70 70
Fam iliarity percent 38 38 37
A nim al percent 11 14 9
D ream er-involved aggression percent 59 51 66
D ream er-involved friendliness percent 70 68 72
A ggressor percent 35 38 32
B efriender percent 52 44 57
Physical aggression percent 45 42 45
A ggression /character index 18 15 21
F riendliness/character index 16 14 18
Settings
Indoor settings percent 43 45 41
Fam iliar settings percent 73 76 70
155
h vs. Male Norms
-1 .0 -0 .8 -0 .6 -0 .4 -0 .2 0 +0.2 +0.4 +0.6 +0.8 +1.0
Male/Female %
Familiarity %
Animal %
Dreamer-Involved Aggression
Aggressor %
Dreamer-Involved Friendliness
Befriender %
Aggresslon/Character Index
Friendliness/Character Index
Indoor Setting %
Familiar Setting %
Figure 6.1. T he h profile fo r the dream er known as “T he Natural S cientist,” using the
m ale norm ative sam ple as a baseline. *p < .05. **p < .01.
dream er-involved friendliness. These findings fit w ith his low -key personality
and his focus on observation in w aking life . In a d d itio n , there are no sexual
in te ra ctio n s in his dreams, w h ich is consistent w ith his status as a life -lo n g
bachelor. O n the other hand, he is s lig h tly elevated on anim al percent,
w h ich fits w ith the fa ct th a t he was raised in a sm all farm tow n, like d to
fish and h u n t, and became an entom ologist. T he anim als in his dreams are

p rim a rily birds, barnyard animals, and the insects he studied in his profes
sional life . He is also somewhat above the norm on fa m ilia r settings percent,
w h ich means, in his case, th a t the dreams are located in his hom e, his
office, and the fa m ily homestead.
The dreams dem onstrate great a tte n tio n to detail, a characteristic th a t
fits w ith the m onum ental taxonom ic task he had undertaken w ith the type
o f insect he studied. T h irte e n dreams refer to golf, w h ich the scientist
enjoyed in w aking life as part o f a life lo n g in vo lve m e n t in com petitive sports.
C o ntrary to w hat m ig h t be expected from activatio n-synth esis theorists’
emphasis on vestibular influences on dream ing, in o n ly one dream is the
scientist flyin g. M ore exactly, in th is dream he is flo a tin g a few yards above
the ground, thanks to a sm all rectangular p la tfo rm th a t he th in ks may be
a carpet and th a t looks somewhat lik e a m agic carpet in the accom panying
draw ing. A s to chase dreams and dreams in w h ich he cannot m ove, w h ich
figure p ro m in e n tly in the examples used by activatio n-synth esis theorists,
he describes none, although in one dream he watches a dog chase a m ountain
lio n . G enerally speaking, then, the connections between dream co n te n t
and his w aking life are more solid th a n any alleged connections between
dream features and his bra in stem.
A ctiva tio n -syn th e sis theorists insist on the need fo r a “w h ole -brain
isom orphism ” in studying dreams, but th e ir isom orphism is prem ature at
best, as evidenced by the m any changes in th e ir specific claims on the basis
o f unexpected neurophysiological findings (Feinberg, 2000; M orriso n &
Sanford, 2000). M oreover, th e ir isom orphism does n o t include any consider
a tio n o f the cogn itive dim ension o f dream ing. T he connections between
b ra in events and dream features th a t are claim ed in th e ir m odel rem ain
e n tire ly hypo the tical. There is no m ore reason to believe, fo r example, th a t
fly in g in dreams is due to vestibular disturbances than there is to believe
th a t it is based on one or another conceptual m etaphor. However, as previous
chapters in this volum e show, it may be possible to dem onstrate em p irica lly
th a t an in te ra ctio n occurs between physiological and cogn itive levels: B rain
variables can affect co g n itive variables, and cogn itive variables can affect
b ra in variables. In fact, as th is book argues, good studies o f dream con te n t
in re la tio n to brain lesions and m edications m ig h t contribu te to a closer
th e o re tica l a rtic u la tio n between b ra in and m ind. Such an a rtic u la tio n w ould
fu lly incorporate findings at the co g n itive le vel and create a neurocognitive
m odel o f dreams.
F U N C T IO N A L TH E O R IE S O F D R EA M S
Innum erable theories o f dream fu n c tio n have been developed (D a lle tt,
1973). A ll o f them are h ig h ly speculative and d iffic u lt to refute in a d e fin itive
way, and they therefore lin g e r despite a la ck o f evidence fo r any o f them .

T h is situ a tio n also provides a fe rtile te rra in fo r new and u n lik e ly theories
th a t are based on analogies draw n from each developm ent or discovery in
o ther areas o f research. T h is search fo r a fu n c tio n seems necessary and
sensible to m ost people, b u t it rests on the false “ adaptationist” assum ption
th a t “ a ll the things th a t have form have fu n c tio n ” (Thom pson, 2000, p.
1014). In fact, m any structures and processes persist even though they have
no fu n ctio n ; dream ing may be one o f them (Flanagan, 1995, 2000a).
A side from Freud’s guardian-of-sleep theory and Jung’s com pensatory
theory, w h ich were refuted earlier in the chapter, the m ost p ro m in e n t theory
o f dream fu n c tio n is th a t dreams provide solutions to curre nt problem s,
especially em otion al problem s (B a rre tt, 1993; G reenberg et al., 1992; G reen
berg &_ Pearlm an, 1993). In one va ria n t, Fiss (1993) suggested th a t dreams
are especially good at registering subtle in te rn a l and external signals th a t
ofte n go undetected in w aking life , m aking them p o te n tia lly useful fo r
p ickin g up early signs o f physical illness.
T here are m any em pirical findings about dreams th a t do n o t fit w e ll
w ith any problem -solving theory. T o begin w ith , the idea th a t dreams have
a purpose originated at a tim e w hen it was th o u g h t th a t people rarely dream.
In th a t co n te xt, it was plausible to believe th a t the occasional recalled
dream could be a re action to a specific event o r em otional problem . B u t if
m ost adults dream at least four to six tim es per n ig h t, th e n m ost people are
recallin g less th a n 1% o f th e ir dreams. Even the best dream recallers rem em
ber o n ly a sm all p ro p o rtio n o f th e ir dreams. T h is la ck o f recall suggests th a t
dreams in general are n o t an evolution ary adaptation to provide in fo rm a tio n
o r in sig h t to people w hen they are awake.
In a d d itio n , o n ly about h a lf o f recalled dreams seem to have even the
slightest con n e ctio n to th e events o f the previous day (B otm an &. C rovitz,
1989; H a rlo w &. R o ll, 1992; H artm ann, 1968; M arquardt et al., 1996; N ielsen
&. Pow ell, 1992). Kram er (2000a) claim ed on the basis o f one sm all c lin ic a l
study th a t the concerns o f the day are incorporated in to dreams, b u t more
recent and larger studies, in w h ich judges try to m atch expressed daytim e
concerns w ith dream reports from laboratory awakenings, have proven unsuc
cessful (Roussy, 2000; Roussy, B runette, et al., 2000). I t is therefore u n lik e ly
th a t dreams o fte n deal w ith im m ediately releva nt issues, although th is book
suggests th a t they dram atize ongoing em otional preoccupations in many
instances.
I f dreams co n ta in im p o rta n t in fo rm a tio n fo r consideration in w aking
consciousness, th e n it m ig h t be predicted th a t people w ho do n o t remember
or pay a tte n tio n to th e ir dreams m ig h t suffer some disadvantages. People
w ho rarely re ca ll dreams, however, do n o t d iffe r in personality o r m ental
d ifficu ltie s from those w ho recall dreams regularly (A ntrobus, 1993; Blagrove
& A ke hurst, 2000; C ohen, 1979; G oodenough, 1991; Tonay, 1993). G ener-
a lly speaking, it is d iffic u lt to distinguish “ recallers” from “ nonrecallers” w ith
e ith e r personality o r cogn itive tests. I f in corp orating and dealing w ith the
c o n te n t o f dreams m attered fo r psychological w ell-being, a d iffe re n t set o f
findings m ight be expected. In fact, contrary to any theory th a t emphasizes
the problem -solving nature o f dreams, dream recall is o fte n as disturbing
as it is helpful, as shown m ost dram atically w ith people w ho suffer from
posttraum atic stress disorder. M any people w ho have recurrent dreams suffer
from them (Zadra, 1996; Zadra &. D onderi, 2000a).
I t is also u n lik e ly th a t dreams co n ta in new in fo rm a tio n on physical
illnesses. The few dream studies cite d by Fiss (1993) are c lin ic a l studies w ith
sm all samples. Those studies cla im to fin d differences in themes concerning
h o s tility or separation, b u t no d ire ct indications o f illness, as Fiss acknow l
edged. His signal-detection theory o f dream fu n c tio n therefore rests on an
extrapolation fro m the lite ra tu re on sublim in al stim ulatio n. A s noted earlier
in th e chapter, m ost research psychologists rem ain h ig h ly d ou btful th a t any
strong effect fo r su b lim in a l stim u la tio n has been dem onstrated or th a t w hat
has been dem onstrated relates to psychodynam ic claim s about the uncon
scious (Fudin, 1999; G reenwald, 1992; G reenwald et al., 1996).
Faced w ith th e findings on the ra rity o f recall and o f dream content
related to current events, some proponents o f problem -solving theories now
cla im th a t o n ly im p o rta n t and strongly em otional dreams have a problem
solving function. There are tw o im p o rta n t d istin ctio n s th a t m ust be made
in analyzing th is claim . First, it is one th in g fo r a dream to “reflect” a
problem ; it is q u ite another fo r it to offer a “ so lu tio n .” Second, a d is tin c tio n
has to be made betw een solutions th a t are present w ith in a dream, on the
one hand, and w aking realizations th a t are based on th in k in g about the
dream , on the other. Realizations in the w aking state are a m uch more
plausible alterna tive because evidence indicates th a t conscious a tte n tio n is
usually needed fo r problem solving (Blagrove, 1992, 1996; Foulkes, 1985).
V iew ed in this way, hum an beings have developed “ uses” fo r dreams in the
course o f history, in clu d in g personal developm ent (Fiss, 1983, 1991; H u n t,
1989); cultural uses, however, are n o t the same as evolved psychological
functions.
C a rtw rig h t’s w ork w ith people going through divorce is sometimes
interpreted as show ing th a t dreams co n ta in solutions, because those who
dream o f the form er spouse “ have a better outcom e” (1996, p. 185). However,
dream ing about a form er spouse does n o t necessarily mean th a t the dreams
co n ta in any solutions to problem s arising from the divorce. Instead, it more
lik e ly simply “reflects” a concern w ith the issue. M oreover, C a rtw rig h t’s study
had severe m ethodological lim its ; her claim s w ould have to be replicated in
new and larger studies before they could be taken seriously as a basis fo r
theorizing. As C a rtw rig h t herself concluded:

The study is suggestive. There are many ways in which it could be
faulted. There was only one night of dream collection, and some who
did n o t dream of the spouse that night m ight well be experiencing a
great deal of incorporation of the problem o f this relationship on other
nights. Also there was a long gap between that one night of dreaming
and the follow-up interview, during which many new reality factors
would have intervened. (Cartwright, 1996, p. 185)
T he d ifficu ltie s o f dem onstrating problem solving in dreams were shown

in a study o f 76 college students between ages 19 and 24. T o increase
relevance and m o tiv a tio n fo r the task, they were allow ed to choose the
problem they hoped to resolve (B arrett, 1993). P articipants were asked to
write o u t th e problem , th in k about it, and keep a dream jo u rn a l fo r a week
or u n til they recalled a dream th a t seemed to solve the problem . B o th the
participants and tw o independent judges rated w hether the dreams were
(a) on the topic and (b ) contained a satisfactory solution.
O n ly h a lf the particip ants recalled a dream they fe lt related to the
problem. T h e dreams usually concerned re la tio n sh ip dilem m as o r educa
tional o r vo ca tio n a l decisions. B o th the dream er and the tw o judges agreed
that in o n ly tw o instances the dreams contained the problem and offered
aplausible solution. B o th dreams seemed to reflect the dreamers’ concerns,
but they d id n o t co n ta in “ solutions.” R ather, they dram atized problem s, as
many dreams do. In the firs t instance the dreamer is “having m ajor problem s
with my m enstrual cycle and my doctor can’t figure out w hat is w rong.”
She reported the fo llo w in g dream, w ith a com m ent about it at the end:
M y doctor told me I was having a reaction from being on a diet and

exercising more than I ever have. In the dream, my doctor gave me
medicine to correct this and I would be fine if I took this medicine. In
waking life, he did ask about diet and I didn’t te ll him how much I ’m
dieting; he’s never asked about exercise. I guess I should te ll him about
diet and exercise, huh? (Barrett, 1993, p. 119)
T h is dream does seem to reflect her concern about n o t te llin g the

whole story to her doctor. B u t the so lu tio n to the problem — te llin g the
doctor about her d ie tin g and exercising— is arrived a t in w aking life by
thinking about the dream scene w herein the doctor tells her w h a t is w rong
and gives her m edicine.
T h e second dream er also has a m edical problem : “ T he problem is
whether I had taken m y m edicine. I ’m supposed to take ju st one o f these
pills a day; it ’s bad i f I take more than one o r miss one. I couldn’t remember
this day if I had taken it and I was really w orried.” She reported the fo llo w in g
dream: “ I was d rin k in g w ater and sw allow ing p ills over and over; it ju st
went on w ith me d rin k in g and taking p ills fo r a long tim e ” (B a rre tt, 1993,
p. 119). O nce again, th is dream seems to reflect an em otional preoccupation,

b u t sw allow ing far m ore p ills than she is supposed to take hardly seems to
be a “ solution” to the problem . Instead, the dream has th a t s lig h tly u n lik e ly
and dram atic q u a lity— w o lfin g dow n a ll those p ills — th a t makes dream-
ing d istin ctive .
In a few cases in B a rrett’s study, the alleged so lu tio n seems to come
during w aking life in reaction to the dream, as Blagrove (1992, 1996) w ould
predict. T h is p o in t is best dem onstrated by a seem ingly m etaphoric dream
th a t supposedly indicated the dreamer should go to graduate school in
Texas or C a lifo rn ia , because “ the lig h t seems to be fu rth e r west” th a n
Massachusetts, her hom e state:
Problem: I have applied to two clinical psychology programs and two

in industrial psychology because I just can’t decide which field I want
to go into. Dream : A map of the United States. I am in a plane flying
over this map. The p ilo t says we are having engine trouble and need
to land and we look for a safe place on the map, indicated by a light.
I ask about Massachusetts, which we seem to be over right then, and
he says all of Massachusetts is very dangerous. The lights seem to be
further west. (Barrett, 1993, p. 118)
The dreamer then reports th a t she “wakes up and realizes th a t my tw o

c lin ic a l schools are b o th in Massachusetts, where I have spent m y w hole
life and where m y parents liv e ,” whereas b o th o f the in d u stria l psychology
programs are far away. She th in ks the dream is te llin g her th a t “ getting
away is more im p o rta n t than w h ich k in d o f program I go to ” (B arre tt, 1993,
p. 118). R ather than the dream te llin g her anything, it is m ore lik e ly th a t
it is portraying w hat she has been th in k in g , because it is lik e ly th a t the
dreamer knew th a t the in d u stria l programs were b o th far from hom e. In
ad d itio n , the realization o f w hat she should do comes to her from w aking
re fle ctio n on the dream.
Thus, little or no systematic evidence supports the hypothesis th a t
dreams in general have a problem -solving fu n ctio n . T he idea therefore is
supported w ith anecdotal testim ony concerning solutions to problem s th a t
presumably arose w ith in dreams. B arrett (2001) assembled a ll the past anec
dotes and adds several new cases th a t are based o n her ow n interview s
w ith architects, artists, scientists, and engineers. Some o f the m ost famous
anecdotes turned out to be bogus or are only know n through secondhand
testim ony; others seem to have emerged during drowsiness, reverie, or drug-
induced states. In one o f the m ost famous examples, the design o f an
experim ent on the transm ission o f nerve impulses, w h ic h led to a N o b e l
Prize in m edicine, the physiologist reports th a t he woke up and w rote dow n
an idea fo r the experim ent; how ever, he does n o t say it came from a dream
(B arre tt, 2001, p. 91). I t is just as lik e ly th a t it came from the th in k in g
th a t can go on during sleep and in b rie f awakenings (A rk in , 1981; Foulkes,

1985). In fact, there are probably m ore everyday examples o f w aking up
w ith a new angle on a problem th a n there are anecdotes about dreams
p ro vid in g a so lu tio n to a problem .
In several o f the cases in w h ich actual dreams were in vo lve d , the
discovery was based on a w aking in te rp re ta tio n o f the dream. For exam ple,
the person w ho invented the process fo r m aking gunshot by dropping m o lte n
lead in to w ater d id so on the basis o f a dream in w h ic h the raindrops p e ltin g
dow n on h im were m olten lead. Based on his know ledge th a t m o lte n lead
forms in to little balls in water, he interpreted the dream to mean th a t
gunshot could be made in this way (B arrett, 2001, p. 113). In a sim ila r
fashion, th e idea fo r how a sewing m achine should h o ld a needle came
from w aking re fle ctio n on a dream in w h ich the n a tive w arriors w ho were
surrounding the dreamer had spears w ith eye-shaped holes near the top.
U p o n aw akening, he decided th a t the hole in the sewing needle also should
be near th e top. N o tab ly, the dream and the in te rp re ta tio n o f it came during
a tim e w hen the in ve n to r was w o rkin g feverishly on his new m achine
(B arrett, 2001, pp. 113 -1 14).
B a rrett also presents cases in w h ic h m usicians have been inspired to
creative efforts by b e a utifu l new m usic they heard in th e ir dreams. In o ther
examples, w riters and poets have used scenes and themes from th e ir dreams
as the basis fo r th e ir w aking w ork. These examples show th a t dreams can
be in sp ira tio n a l and provide the basis fo r new creative efforts in w aking
life , b u t th a t is n o t the same th in g as solving a problem . A s Foulkes (1999)
argued, every dream is a n o ve l and creative construction, b u t th a t does n o t
mean dreams necessarily have a problem -solving fu n ctio n . Instead, the
examples dem onstrate w aking consciousness m aking use o f dream content.
W h e n a ll is said and done, then, o n ly occasional anecdotal evidence
supports the idea th a t dream ing its e lf provides any solutions to problem s.
T h is anecdotal evidence is n o t im pressive w hen it is seen in the co n te xt
o f the sm all percentage o f dreams th a t are recalled and the even sm aller
percentage o f recalled dreams th a t m ig h t be construed as having a so lu tio n
to a problem . Dreams sometimes can be useful to w aking consciousness as
a basis fo r th in k in g about problem s in a new way, o r as a basis fo r discussing
personal problem s, as c lin ic a l research shows (Fiss, 1991; G reenberg et al.,
1992; H ill, 1996). Dreams th a t have a dram atic em otion al im pact create a
strong subjective sense th a t they m ust have an im p o rta n t message. However,
it does n o t fo llo w from usefulness o r a w aking im pression o f p ro fu n d ity th a t
dream ing has an adaptive fu n c tio n (A ntrobus, 1993).
D o U n re ca lle d Dream s H ave a F unction?
Recognizing the problem s w ith theories th a t a ttrib u te a w aking fu n c

tio n to the few dreams th a t are recalled, several theorists now cla im th a t

dream ing has a fu n c tio n even when dreams are n o t recalled. D raw ing on
an analogy w ith com puters, one pair o f theorists claim s th a t dream ing clears
o u t useless memories from the day before (C ric k & M itch iso n , 1983, 1986).
D raw ing on speculations in evolutionary psychology, others cla im th a t
dreams have a social learning (Brereton, 2000) or thre a t-sim u la tio n fu n c tio n
(Revonsuo, 2000). S till others a ttrib u te psychodynam ic functions to unre
called dreams (H artm ann, 1998; Kram er, 1993).
T he idea th a t dreams help rid th e b ra in o f useless recent memories is
based on the assum ption th a t the p o n tin e instigators o f REM are producing
random , meaningless imagery. T h is claim depends cru cia lly on the contested
cla im th a t dream ing is confined to REM , b u t it is also challenged by three
o th e r findings. First, o n ly about h a lf o f recalled dreams have even one slig h t
reference to recent events, w h ich is contrary to w hat the theory w ould
p re d ict (B otm an & C rovitz, 1989; H a rlow & R o ll, 1992). Second, dreams
are more coherent and related to w aking thoughts than the theory w ould
predict. T h ird , the findings on the consistency o f dream content over years
and decades, especially negative dream content, is opposite o f w h at this
theory w ould predict; no evidence supports the idea th a t re p e titive dream
co n te n t clears o u t useless memories (D om hoff, 1996; see also chapter 5).
T he rehearsal theories put fo rth by B rereton (2000) and Revonsuo
(2000) attem pt to b u ild on the evidence fo r the consolidation o f procedural
memories during R E M (S m ith, 1995), but m em ory consolidation is n o t the
same th in g as new learning during sleep thro u g h m ental rehearsal, fo r w h ich
there is no evidence. Even though some m em ory consolidation occurs during
sleep, it does n o t fo llo w th a t dream ing is also occurring (A ntrobus, 1993;
Flanagan, 2000b).
T he rehearsal theories assume th a t R E M and dream ing are one and
the same in a ll mammals, w h ich is a dubious assumption. T he y use the
evidence fo r m em ory consolidation during sleep in anim al studies to support
th is assumption, but there is a strong case th a t no other animals, even other
prim ates, have the conceptual capacities to dream (Foulkes, 1983). T he
m ost notable cla im fo r anim al dream ing is based on the exploratory and
aggressive behaviors displayed by sleeping cats w ith experim ental lesions in
the area o f the b ra in stem th a t in h ib its m ovem ent during REM (Jouvet,
1999, chapter 4). However, as Foulkes (1983) argued, behavior d urin g sleep
does n o t necessarily in dica te dream ing, as seen first o f a ll by the fa ct th a t
sleepw alking starts in the deepest stages o f N R E M and usually does n o t lead
to dream reports w hen awakened. M oreover, the same m ovem ents observed
durin g sleep in the cats w ith experim ental lesions also occur perio d ica lly
w hen they are awake.
Advocates o f anim al dream ing sometimes p o in t to an unpublished
dissertation in w h ich monkeys were trained to press a bar to avoid w aking
visual images. A fte r they were trained, they were fitte d w ith contact lenses

to e lim in a te patterned visual stim u la tio n , and then supposedly, and q u ite
unexpectedly, e xh ib ite d the bar pressing response during sleep (Vaughan,
1963). H pw ever, there are numerous m ethodological problem s w ith the
study. Because the study actually was concerned w ith the effects o f sensory
d e p riva tio n during the w aking state, n o t sleep, no electrophysiological re
cordings were made to ve rify sleep, an extrem ely seriously problem in th a t
the monkeys seemed to be comatose fo r the firs t 27 hours after the lenses
were placed over th e ir eyes and la te r w ould rem ain m otionless fo r hours at
a tim e (Foulkes, 1983; Vaughan, 1963, pp. 86, 93, 103). M oreover, the fact
th a t the bar pressing episodes occurred m any hours apart and lasted fo r as
long as 5 hours does n o t fit w ith the p e rio d ic ity o f REM periods or th e ir
usual length. I t is a dam ning com m entary th a t such a seem ingly im p o rta n t
fin d in g was never published or replicated. I t therefore cannot be given any
c re d ib ility in terms o f anim al dream ing.
T he in trig u in g idea th a t o n ly hum an beings dream is greeted w ith
great surprise and im m ediate re je ctio n by m ost dream theorists. T he y are
n o t impressed w ith the argum ent th a t the la ck o f co g n itive skills fo r dream ing
in young c h ild re n suggests th a t less developed anim als probably do n o t
dream, and they re je ct the m ore d ire ct argum ents in the previous paragraphs
against the lik e lih o o d th a t e ith e r cats or m onkeys dream. Instead, they say
th a t Foulkes’s evidence w ith preschool c h ild re n and his arguments concern
in g o ther anim als are n o t foolp ro of, m eanw hile ignoring the fa ct th a t th e ir
ow n positive claim s about dream ing are far less plausible and fa r less lik e ly
to be supported. Thus, they im p lic itly seem to recognize th a t the stakes are
great here. I f on ly hum an beings over age 5 o r 6 are able to dream, th e n
evo lu tio n a ry theories o f dream fu n c tio n lose a ll c re d ib ility .
A n im a l dream ing aside, there are fu rth e r problem s w ith Revonsuo’s
version o f the rehearsal theory. I t stretches the im aginatio n to th in k th a t
the o n e -tria l system o f fear c o n d itio n in g th a t has been present in the b ra in
since the e v o lu tio n o f reptiles needs to be prim ed by dream ing (LeD oux,
1996). T he low levels o f dream ing in young ch ild re n and the benign nature
o f the few dreams they do have do n o t support Revonsuo’s cla im th a t dreams
are useful in h elping c h ild re n learn to be v ig ila n t against dangerous anim als
(Foulkes, 1999). H is fu rth e r cla im th a t traum a may stim ulate dream ing in
c h ild re n does n o t seem plausible in the lig h t o f Foulkes’s (1982) fin d in g
th a t c h ild re n w ith tense hom e environm ents d id n o t report m ore dreams
th a n o ther ch ild re n . H is visio n o f the “ ancestral enviro n m e n t” as being
fille d w ith dangerous predators does n o t seem credible in the lig h t o f prim ate
evidence th a t fe llo w members o f th e ir ow n group are the biggest threat.
Revonsuo’s theory does n o t e xplain the large am ount o f dream co n te n t
th a t does n o t relate to th re a t and aggression (Zadra & D onderi, 2000b).
H e overstates the am ount o f physical aggression in dreams by dow nplaying
the d is tin c tio n between physical and nonphysical aggression in the H a ll-
V an de Castle coding system. F in a lly, the theory ignores the fact th a t
nightm a rish attack dreams are d e b ilita tin g fo r m any people, m aking them
less f it fo r d aily life (Levin , 2000; N ielsen & G erm ain, 2000).
F u n c tio n a lis t Psychodynam ic T heories
B u ild in g on the general psychodynam ic idea th a t dream ing is an at-

tem pt to deal w ith personal problem s, H artm ann (1998) argued th a t the
fu n c tio n o f dreams is to help people w o rk through trau m atic experiences,
w hether th e dreams are recalled or n o t. A cco rd in g to his theory, dreams
deal w ith a traum a by p u ttin g it in m any d iffe re n t m ental contexts w ith in
the “ safe place” o f sleep, where psychological “ connections” can be made
w ith o u t any personal danger. Studies show ing th a t free associations are more
u n lik e ly (i.e., “ distan t” ) and im agistic in the w aking period shortly after
REM than after N R E M lend support to this aspect o f the theory (Fiss et
al., 1966, 1969; S tickgold, S cott, R ittenhouse, & Hobson, 1999).
A lth o u g h H artm ann’s theory does n o t require th a t dreams be recalled,
the evidence fo r it rests on alleged changes in th e dreams th a t traum a
victim s happen to recall. First, H artm ann assumes th a t the recalled dreams
reflect the trauma, often in d ire ctly, through w hat he calls a contextua lizing
m etaphor. These s trikin g images, such as the approach o f a tid a l wave,
epitom ize the way in w h ich these dreams are fille d w ith in tim a tio n s o f
disaster, m isfortunes, violence, and strong negative em otions. Second, he
asserts th a t the posttraum a dreams become less negative over tim e, suggesting
th a t the traum a is being dealt w ith by the dream process.
However, th e evidence fo r b o th o f these claim s is weak. A lth o u g h
the dream examples H artm ann (1998) presented are fille d w ith negative
em otions, his coding system fo r em otions is unproven, and he offered no
norm ative data to show th a t the dreams are, in fact, m ore negative than
the dreams o f those who have n o t suffered traumas (see D om hoff, 1999c,
fo r a fu ll critiq u e o f H artm ann’s m ethod and evidence). For example, if the
negative em otions percent in the H a ll-V a n de Castle norm s is taken as a
baseline, then it can be expected th a t 80% o f the em otions in any dream
sample w ill be negative. The same figure was reported in three la te r studies
th a t used dream reports from d iffe re n t eras, the sleep laboratory, and Canada
(H a ll et ah, 1982; Roussy, Raym ond, & De K o n in ck, 2000; Tonay, 1990/
1991).
H artm ann (1998) also noted the large am ount o f aggression and other
negative events in the dreams o f traum a victim s, b u t he did n o t provide
any system atic studies. I f H a ll-V a n de Castle norm ative findings are once
again taken as a baseline, th e n 23% o f m en’s dreams and 15% o f wom en’s
dreams have at least one chase, attack, o r m urder, and 33% o f men’s dreams
and 36% o f wom en’s dreams have a t least one m isfortune. Even more to

the p o in t o f H a rtm a nn’s studies, w h ich used five recent dreams from each
traum a v ic tim , 69% o f the m en and 54% o f the w om en in the H a ll-V a n
de C astle norm ative sample had at least one chase, attack, o r m urder in
the five dreams they contribu ted .
H a rtm a n n ’s theory cru cia lly hinges on the degree to w h ich the dreams
o f traum a victim s change over tim e. H e asserted (1998) th a t he found
p o sitive changes in several dream series, b u t he presented no system atic
data o n declines in p a rticu la r negative themes or in any con te n t categories
th a t p a ra lle l the categories fo r aggressions, m isfortunes, failures, and negative
em otions in the H a ll-V a n de Castle system. H e also has n o t reported any
com parisons w ith possible changes in the dreams o f people w ho have n o t
suffered traumas. T he general consistency o f dreams over m onths and decades
th a t has been dem onstrated in th is book, along w ith the po ssib ility th a t
the rate o f aggression in dreams may decline w ith age, provides reason to
be skeptical about his claim s u n til m ore system atic findings w ith an adequate
coding system are presented. Such studies m ig h t show th a t the dream co n te n t
a ctually stays th e same b u t th a t the dreamer is less upset in the face o f it.
T h is o p tio n is n o t b u ilt in to H a rtm a n n ’s theory, b u t it is suggested by the
findings on the H ow ard subseries in chapter 5. Ideally, future tests o f the
theory w ould include dreams from before th e trau m atic event to provide
the best possible baseline.
A s do o th e r c lin ic a lly derived theories o f dream ing, H a rtm a n n ’s theory
assumes th a t the conceptual m etaphors so im p o rta n t in w aking th o u g h t are
also operating in dreams. H ow ever, as pointed o u t m any tim es in th e course
o f th is book, and as shown at the end o f chapter 5, there is s till no solid
evidence to support th is claim . Even if H a rtm a nn were able to answer
the e m p irica l objections discussed in the three previous paragraphs, the
m etaphoric basis o f his theory w ould rem ain an untested assumption.
In a revision o f Freud’s theory o f dreams as the guardians o f sleep,
K ram er (1993) developed a theory th a t takes in to account the regularity
o f dream ing during REM . H e argued th a t the fu n c tio n o f dreams is to
co n ta in an “ em otion al surge” th a t builds up in the course o f each R EM
period, an idea suggesting th a t th e dream process is successful if the person
does n o t wake up and re ca ll the dream. H e believes th a t the larger percentage
o f R E M tow ard m orning is consistent w ith this idea because the urge to
awaken is greater at th a t p o in t. Kram er presented suggestive evidence from
a w ide range o f his ow n studies, some on ly published in abstract form o r
labeled as p ilo t studies, to make a case th a t dreams are “ responsive to a
num ber o f affective influences” (1993, p. 175).
T h is theory has several problem s. First, it does n o t account fo r N R E M
dream ing and cannot e xp la in w hy young c h ild re n and some adults sleep
w e ll enough w ith o u t dreams. Furtherm ore, the evidence does n o t cons is-

te n tly suggest the presence o f a grow ing em otional surge during R EM .
Instead, it suggests a rise in in te n sity in the m iddle o f the REM period, as
indexed by the density o f eye movements, follow e d by a gradual decline.
N o r does there seem to be an accum ulation o f undischarged em otional drive
w hen people are deprived o f REM , u n like w h at m ig h t be expected on the
basis o f Kram er’s theory (Greenberg & Pearlman, 1993, 1999).
As one piece o f evidence fo r his theory, Kram er claim ed th a t dream
co n te n t changes from R E M period to R E M period throughout the n ig h t,
p a rticu la rly in the num ber o f characters, b u t m ost studies have found few
o r n o differences (Foulkes, 1966,1985; H a ll, 1966b). A cco rding to Kram er’s
theory, recallin g a dream “ depends to a degree on a troubled state in the
dreamer” (Kram er, 1993, p. 187), b u t studies o f h ig h and lo w dream recallers
do n o t lend any support to th is cla im (Blagrove &. A kehurst, 2000; Tonay,
1993). O ne study showed th a t wom en recall m ore dreams w hen under stress
b u t th a t m en recall fewer dreams (A rm itage, 1992). In add ition , as Kram er
h im se lf says (1993), “ Efforts to connect co n te n t w ith v a ria b ility in autonom ic
variables have been m in im a lly successful” (p. 145). M oreover, even if dreams
are responsive to affective influences, it does n o t necessarily fo llo w th a t
th e ir fu n c tio n is to co n ta in em otions during sleep.
I f th e weaknesses o f a ll theories o f dream fu n c tio n are com bined w ith
the evidence th a t dream ing is a process th a t occurs only in hum an beings
and on ly after age 5 o r 6, th e n it seems h ig h ly u n lik e ly th a t dreams have
any adaptive fu n ctio n . They cu rre n tly seem to be the by-product o f tw o
great evolutionary developm ents, sleep and com plex co g n itive processes.
These m in im a list conclusions jo in w ith em pirical findings presented
thro ugho ut th is book to suggest th a t co g n itive psychology may be the best
startin g p o in t fo r developing an adequate m odel o f dreams.
C O N C L U S IO N
As th is chapter shows, a large am ount o f system atic em pirical evidence

does n o t fit w ith the c lin ic a l, neuropsychological, and fu n c tio n a l theories
o f dream ing th a t predom inated in the 20th century. T h is evidence, how ever,
is com patible w ith an open-ended neurocognitive m odel. T h is new m odel
stresses the ways in w h ic h a neural netw o rk fo r dream ing can be integrated
w ith findings from developm ental studies o f dream ing and the con te n t
analysis o f dream reports.
T he m odel offers hypotheses concerning a ll the key questions th a t
have been asked about dream ing in the past. I t first addresses the issue o f
the in stig a tio n o f dream ing by suggesting th a t th is cogn itive process occurs
167
in m ost a d u lt m inds w hen a ce rta in m in im a l le ve l o f neural a c tiv a tio n
occurs in th e c o n te xt o f an occlusion o f external s tim u li and a re lin q u ish m e n t
o f self. T h e m odel therefore can account fo r dream ing at sleep onset, in
b o th R E M and N R E M sleep, and even durin g relaxed w aking states.
By specifying a neural netw ork fo r dream ing th a t can develop defects,
the neu ro cognitive m odel is able to e xp la in the absence o f dream ing in
some adults, such as those who have suffered b ra in in jurie s in specific
lo calities, as w e ll as excesses o f dream ing caused by b ra in in ju rie s or
neurochem ical imbalances. A lth o u g h it has a neural grounding, the m odel
also suggests th a t dream ing is a co g n itive achievem ent th a t depends on
the developm ent o f a range o f co g n itive skills, especially the visuospatial
skills th a t make m ental imagery possible. T h is developm ental dim ension
explains w hy preschool ch ild re n rarely report dreams after laboratory
awakenings as w e ll as w hy the dreams they do report are b rie f and static
in nature.
By ta kin g seriously the in fo rm a tio n on nondream ing in ch ild re n and
adults w ith b ra in lesions, the neurocognitive m odel is able to approach the
question o f dream fu n c tio n by suggesting th a t dream ing may have no func-
tio n . T he m odel raises the p o ssibility th a t dream ing is a spandrel o f the
m ind, a by-product o f the e vo lu tio n o f sleep and consciousness. A lth o u g h
dreams probably have no fu n ctio n , evidence suggests th a t they have at least
some coherence and m eaning. Thus, the tendency to conflate fu n c tio n and
m eaning is n o t present in the neurocognitive m odel.
Laboratory awakenings show th a t m ost dreams are reasonable sim ula
tions o f the w aking w o rld in habite d by the dreamer. T he co n te n t analysis
o f dreams fro m people o f a ll ages from m any d iffe re n t parts o f the w o rld
suggests th a t dreams ofte n express conceptions and concerns by at least
preadolescence. I t is therefore lik e ly th a t they use m any o f the same schemata
and scripts th a t are available to w aking th o u g h t. These parallels w ith w aking
th o u g h t e xp la in w hy dreams can be useful in psychotherapy. H ow ever, it
may be th a t dream ing is less constrained by present re a lity in m aking use
o f these schemata— w h ich is one reason th a t dreams can strike people
as bizarre.
By stressing the sim ilarities between dream ing and w aking co g n itio n ,
the neu ro cognitive m odel opens up the p o ssib ility th a t some o f the more
puzzling aspects o f dream co n te n t may be a product o f the system o f figurative
th o u g h t th a t is so pervasive in w aking life . T he processes o f m etaphor,
m etonym y, iro n y, and conceptual blending may be the germ o f tru th in
Freud’s claim s about the dream -w ork. T he parallels th a t dreams have w ith
w aking fig u ra tive th o u g h t can be used to e xp la in w hy dreams have religious
and m e d icin a l uses th a t were invented by people in d iffe re n t cultures in
the course o f history.
Despite the p o ssib ility th a t figurative th in k in g may be present in some
dreams, the m odel does n o t assume th a t every aspect o f every dream is
somehow psychologically m eaningful. T he e xte n t o f m eaning is an open
question th a t can on ly be answered through better and more detailed searches
fo r dream m eaning. T he degree o f coherence and meaningfulness in dreams
is an em pirical question th a t m ust be studied in great d e ta il before any
conclusions can be reached. T h is book shows th a t psychological in fo rm a tio n
can be extracted from dream reports, a fin d in g th a t im plies th a t dreams
have some m eaning, b u t it also stresses th a t m uch dream co n te n t is s till
n o t understood and may tu rn out to be the product o f freew heeling im provisa
tio n o f little im port.
The neurocognitive m odel has the v irtu e o f being e m in e n tly testable
in a variety o f ways using the neuroim aging technologies and software
programs fo r co n te n t analysis th a t became available in the 1990s. There is
reason to believe th a t the H a ll-V a n de Castle coding system may be o f
value in th is effo rt, especially w hen applied to long-term dream journals
supplied by people who have developm ental anomalies, b ra in injuries, or
psychic traumas or who are taking a dream -enhancing or dream-suppressing
m edication. R ating scales fo r e m o tio n a lity-e va lu a tio n , ra tio n a lity -
bizarreness, a c tiv ity , and im pression-vividness also can be used. I t m ig h t
even be possible to use the search program and dream archive on Dream -
B ank.net to conduct systematic studies o f “ sym bolism ” by testing some o f
the ideas on figurative th in k in g th a t have been developed by co g n itive
linguists and psycholinguists. As stressed in chapter 1, however, m etaphorical
interpretations o f dreams m ust be seen as the fo o l’s gold o f dream theories
u n til systematic evidence fo r them is produced.
The neurocognitive m odel is n o t bogged dow n by arguments about
w hether the forebrain dream ne tw o rk always requires stim u la tio n from the
R E M generator in the pons in order to fu n ctio n . Even i f it turns out th a t
th is area is always the a ctiva tin g source fo r the neural netw o rk fo r dream ing,
the im p o rta n t p o in t is th a t the conceptual systems in the forebrain p o rtio n
o f the netw ork produce dreams. M oreover, the theory does n o t concern
its e lf w ith the exact nature o f the neurom odulation during dream ing. Instead,
it starts w ith the phenom enon o f dream ing its e lf and the dream reports th a t
sometimes eventuate from dream ing, and it th e n attem pts to see how they
m ig h t relate to the neuropsychological level. I t is a theory th a t puts dreams
back in to the fie ld o f dream research and attem pts to relate dream ing to
findings in co g n itive psychology as w e ll as neurophysiology.
Dreams are m uch m ore coherent and m eaningful th a n has been claim ed
by those who overlook the co g n itive dim ension o f dream ing and instead
focus on the neurophysiological and neuropsychological levels. A t the same
tim e, dreams also seem to be m uch less profound than claim ed by Freud o r

Jung. In lig h t o f th e system atic dream research used to develop the new
m odel o f dream ing and dreams described in th is volum e, it now seems lik e ly
th a t dream ing is a com prehensible co g n itive process w ith m any sim ila ritie s
to w aking though t. I t rem ains fo r futu re co g n itive scientists to test, amend,
and expand th is neu ro cognitive m odel so th a t the la te -n ig h t m ovies in the
b ra in can be incorporated in to am bitious theories seeking to e xp la in a ll
aspects o f the hum an m ind.

REFERENCES
Abel, B. (1994). The dream content o f menopausal women: A n exploratory study.

Unpublished doctoral dissertation, U nited States International University,
San Diego, CA.
Abel, V. (1996). Cognitive processing styles o f creative artists as observed in dream
m entation and waking measures. Unpublished doctoral dissertation, Yeshiva
University, New York.
Adolphs, R., & Damasio, A . (1998). The human amygdala in social judgment.
N ature, 393, 470-474.
A jilore, O., Stickgold, R., Rittenhouse, C., & Hobson, J. (1995). Nightcap: Labora-
tory and home-based evaluation o f a portable sleep m onitor. Psychophysiology,
32(1), 92-98.
A llp o rt, G. (1942). The use o f personal documents in psychological science. New York:
Social Science Research Council.
Antrobus, J. (1978). Dreaming as cognition. In A . A rkin , J. Antrobus, & S. Ellman
(Eds.), The m ind in sleep: Psychology and psychophysiology (pp. 569-581). H ills
dale, NJ: Lawrence Erlbaum.
Antrobus, J. (1983). REM and NREM sleep reports: Comparisons of word frequen
cies by cognitive classes. Psychophysiology, 20, 562-568.
Antrobus, J. (1991). Dreaming: Cognitive processes during cortical activation and
high afferent thresholds. Psychological Review, 98, 96-121.
Antrobus, J. (1993). Dreaming: Could we do w ithout it? In A . M offitt, M. Kramer,
& R. Hoffmann (Eds.), The functions o f dreaming (pp. 549-558). Albany: State
University of New York Press.
Antrobus, J. (2000a). How does the dreaming brain explain the dreaming mind?
Behavioral and Brain Sciences, 23, 904-907.
Antrobus, J. (2000b). Theories of dreaming. In M . Kryger, T. Roth, & W . Dement
(Eds.), Principles and practices o f sleep medicine (3rd ed., pp. 472-481). Philadel
phia: W . B. Saunders.
171
Antrobus, J., Kondo, T., & Reinsel, R. (1995). Summation of REM and diurnal
cortical activation. Consciousness and C ognition, 4 , 275-299.
A rkin , A . (1981). Sleep talking: Psychology and psychophysiology. Hillsdale, NJ:
Erlbaum.
A rkin , A ., & Antrobus, J. (1991). The effects of external stim uli applied prior to
and during sleep on sleep experience. In S. Ellman & J. Antrobus (Ed.), The
mind in sleep: Psychology and psychophysiology (2nd ed., pp. 265-307). New
York: W iley & Sons.
Armitage, R. (1992). Gender differences and the effect of stress on dream recall:
A 30-day diary report. Dreaming, 2, 137-142.
Armitage, R., Rochlen, A ., Fitch, T., Trivedi, M., & Rush, A . (1995). Dream recall
and major depression: A prelim inary report. Dreaming, 5, 189-198.
Aserinsky, E., & Kleitm an, N . (1953). Regularly occurring periods of eye m otility,
and concomitant phenomena, during sleep. Science, 118, 273-274.
Auld, F., Goldenberg, G., & Weiss, J. (1968). Measurement of primary process
thinking in dream reports. Journal o f Personality and Social Psychology, 8, 418-
426.
A vila-W hite, D., Schneider, A., & Domhoff, G. W . (1999). The most recent
dreams o f 1 2 -13-year-old boys and girls: A methodological contribution to
the study of dream content in teenagers. Dreaming, 9, 163-171.
Ayella, M . F. (1998). Insane therapy: P ortrait o f a psychotherapy cult. Philadelphia:
Temple University Press.
Baekland, F., & Lasky, R. (1968). The morning recall of rapid eye movement period
reports given earlier in the night. Journal o f Nervous and M ental Diseases,
147, 570-579.
Baldwin, A . (1942). Personal structure analysis: A statistical method for investigat
ing the single personality. Journal o f Abnorm al and Social Psychology, 37, 163-
183.
Bancaud, J., Brunet-Bourgin, F., Chauvel, P., & Halgren, E. (1994). Anatom ical
origin o f deja vu and vivid “ memories” in human temporal lobe epilepsy. Brain,
117, 71-90.
Barrett, D. (1992). Just how lucid are lucid dreams? Dream ing, 2, 221—228.
Barrett, D. (1993). The “committee o f sleep” : A study o f dream incubation for
problem solving. Dream ing, 3, 115-122.
Barrett, D. (1996). Traum a and dreams. Cambridge, M A : Harvard U niversity Press.
Barrett, D. (2001). The committee o f sleep: New York: Crown/Random House.
Barrett, D., & Loeffler, M . (1992). Comparison o f dream content o f depressed vs.
nondepressed dreamers. Psychological Reports, 70, 403-406.
Beck, A ., & Hurvich, M . S. (1959). Psychological correlates o f depression: I.
Frequency of “ masochistic” dream content in a private practice sampling.
Psychosomatic M edicine, 21, 5 0 -5 5 .
Bednar, J. A . (2000). Internally-generated activity, non-episodic memory, and
em otional salience in sleep. Behavioral and Brain Sciences, 23, 908-909.
Belicki, K. (1987). Recalling dreams: A n examination of daily variation and in divid
ual differences. In J. Gackenbach (Ed.), Sleep and dreams: A sourcebook (pp.
187-206). New York: Garland.
Bell, A ., & Hall, C. (1971). The personality o f a child molester: A n analysis o f dreams.
Chicago: Aldine.
Berger, R. (1967). W hen is a dream is a dream is a dream? Experimental Neurology,
4 ( SuppL), 15-27.
Berger, R. (1969). The sleep and dream cycle. In A . Kales (Ed.), Sleep: Physiology
and Pathology (pp. 17-32). Philadelphia: Lippincott.
Bernstein, D., & Belicki, C. (1995). Assessing dreams through self-report question
naires: Relations w ith past research and personality. Dreaming, 5, 13-27.
Berrien, F. (1933). A statistical study of dreams in relation to emotional stability.
Journal o f Abnorm al and Social Psychology, 28, 194-197.
Blagrove, M. (1992). Dreams as a reflection o f our waking concerns and abilities:
A critique o f the problem-solving paradigm in dream research. Dreaming,
2, 205-220.
Blagrove, M. (1996). Problems w ith the cognitive psychological modeling o f dream
ing. Journal o f M in d and Behavior, 17, 99-134.
Blagrove, M., & Akehurst, L. (2000). Personality and dream recall frequency:
Further negative findings. Dreaming, 10, 139-148.
Block, H. (1999, June). The ex-husband dreams in along dream journal. Paper presented
at the annual meeting o f the Association for the Study o f Dreams, Santa
Cruz, CA.
Bonato, R. A., M ofhtt, A ., Hoffmann, R., & Cuddy, M. A . (1991). Bizarreness in
dreams and nightmares. Dreaming, 1(1), 53-61.
Boselli, M., Parrino, L., Smerieri, A., & Terzano, M. (1998). Effect o f age on EEG
arousal in normal sleep. Sleep, 21, 351-357.
Bosinelli, M., & Cicogna, P. C. (2000). REM and NREM mentation: Nielsen’s
model once again supports the supremacy o f REM. Behavioral and B rain Sciences,
23, 913-914.
Boss, M. (1958). The analysis o f dreams. New York: Philosophical Library.
Boss, M. (1977). I dreamt last night. New York: Gardner Press.
Botman, H., & Crovitz, H. (1989). Dream reports and autobiographical memory.
Im agination, C ognition and Personality, 9, 213-214.
Bowen, M., & Dunn, S. (1999, June). Findings from a long dream journal. Paper
presented at the annual meeting of the Association for the Study of Dreams,
Santa Cruz, C A.
Bradley, L., H ollifield, M., &. Foulkes, D. (1992). Reflection during REM dreaming.
Dreaming, 2, 161-166.
Braun, A ., Balkin, T., Wesensten, N., Carson, R., Varga, M ., Baldwin, P., et al.
(1997). Regional cerebral blood flow throughout the sleep-wake cycle: A n
(H 20)-0-15 PET study. Brain, 120, 1173-1197.
REFERENCES 173
Braun, A ., Balkin, T., Wesensten, N., Gwadry, F., Carson, R., Varga, M ., et al.
(1998). Dissociated pattern of activity in visual cortices and their projections
during human rapid eye movement sleep. Science, 279, 91-95.
Brenneis, C. (1975). Developmental aspects of aging in women: A comparative
study of dreams. Archives o f General Psychiatry, 32, 429-434.
Brereton, D. (2000). Dreaming, adaptation, and consciousness: The social mapping
hypothesis. Ethos, 28, 379-409.
Buckley, J. (1970). The dreams o f young adults. Unpublished doctoral dissertation,
Wayne State University, Detroit, M I.
Bulkeley, K. (1999). Visions o f the night: Dreams, religion, and psychology. Albany:
State U niversity of New York Press.
Bulkeley, K. (Ed.). (2001). Dreams: A reader on the religious, cultural, and psychological
dimensions o f dreaming. New York: Palgrave.
Bulkeley, K., Dunn, S., & Domhoff, G. W . (2001). Creating categories fo r the study
o f im pactful and memorable dreams. Unpublished manuscript, Santa Clara U n i
versity.
Bursik, K. (1998). M oving beyond gender differences: Gender role comparisons of
manifest dream content. Sex Roles, 38, 203-214.
Busink, R., &. Kuiken, D. (1996). Identifying types o f im pactful dreams: A replica
tion. Dream ing, 6, 97-119.
Butler, S., & Watson, R. (1985). Individual differences in memory for dreams: The
role of cognitive skills. Perceptual and M o to r Skills, 53, 841-864.
Calkins, M . (1893). Statistics o f dreams. A merican Journal o f Psychology, 5 , 311-343.
Cartwright, D. (1953). Analysis of qualitative material. In L. Festinger & D. Katz
(Eds.), Research methods in the behavioral sciences (pp. 421-470). New York:
H olt, Rinehart, and W inston.
Cartwright, R. (1992). Masochism in dreaming and its relation to depression.
Dreaming, 2, 79-84.
Cartwright, R. (1996). Dreams and adaptation to divorce. In D. Barrett (Ed.),
Traum a and dreams (pp. 179-185). Cambridge, M A : Harvard U niversity Press.
Cartwright, R., & Romanek, I. (1978). Repetitive dreams of normal subjects. Sleep
Research, 7, 174.
Cavallero, C., & Foulkes, D. (1993). Dreaming as cognition. New York: Harvester
Wheatsheaf.
Ceci, S. J., Bruck, M ., & Battin, D. B. (2000). The suggestibility of children’s
testimony. In D. F. Bjorklund (Ed.), False^memory creation in children and
adults: Theory, research, and implications (pp. 169-201). Mahwah, NJ: Lawrence
Erlbaum Associates.
Chapman, P., & Underwood, G. (2000). M ental states during dreaming and day
dreaming: Some methodological loopholes. Behavioral and B rain Sciences,
23, 917-918.
Chugani, H . (1999). M etabolic imaging: A window on brain development and
plasticity. Neuroscientist, 5, 29-40.
REFERENCES
Cicogna, P., Natale, V., Occhionero, M ., & Bosinelli, M. (1998). A comparison o f
mental activity during sleep onset and morning awakening. Sleep, 2 1 , 462-470.
C ipolli, C. (2000). Iterative processing o f information during sleep may improve
consolidation. Behavioral and B rain Sciences, 23, 919.
C ipolli, C., & Poli, D. (1992). Story structure in verbal reports o f mental sleep
experience after awakening in REM sleep. Sleep, 15, 133-142.
Clark, J., Trinder, J., Kramer, M ., Roth, T., & Day, N. (1972). A n approach to
the content analysis o f dream scales. In M. Chase, W . Stern, & P. W alter
(Eds.), Sleep research (Vol. 1, pp. 118-119). Los Angeles: Brain Research
Institute, University of California.
Cohen, D. (1979). Sleep and dreaming. New York: Pergamon Press.
Cohen, D., & W olfe, G. (1973). Dream recall and repression: Evidence for an
alternative hypothesis. Journal o f Consulting and C linical Psychology, 4 1 , 349-
355.
Cohen, J. (1977). Statistical power fo r the behavioral sciences. New York: Academic
Press.
Cohen, J. (1990). Things I have learned (so far). Am erican Psychologist, 45, 1304-
1312.
Cohen, J. (1994). The earth is round (p < .05). Am erican Psychologist, 49, 997-1003.
Conduit, R., Crewther, S. G., & Coleman, G. (2000). Shedding old assumptions and
consolidating what we know: Toward an attention-based model o f dreaming.
Behavioral and Brain Sciences, 23, 924-928.
Cory, T. L., Ormiston, D. W ., Simmel, E., & Dainoff, M . (1975). Predicting the
frequency o f dream recall. Journal o f Abnorm al Psychology, 84, 261-266.
Cote, L., Lortie-Lussier, M ., Roy, M ., & DeKoninck, J. (1996). The dreams of
women throughout adulthood. Dreaming, 6, 187-199.
Crick, F., & M itchison, G. (1983). The function o f dream sleep. N ature, 304, 111-
114.
Crick, F., & M itchison, G. (1986). REM sleep and neural nets. Journal o f M ind
and Behavior, 7, 229-250.
Dallett, J. (1973). Theories of dream function. Psychological B ulletin, 79, 408-416.
Damasio, A . (1999). The feeling o f what happens. New York: Harcourt Brace &
Company.
Damasio, A ., Graff-Radford, N ., Eslinger, P., Damasio, H ., & Kassell, N . (1985).
Amnesia following basal forebrain lesions. Archives o f Neurology, 42, 263-271.
de Rivera, J., & Sarbin, T. (Eds.). (1998). Believeddn imagings: The narrative construe-
tion o f reality. Washington, DC: American Psychological Association.
Dement, W . (1955). Dream recall and eye movements during sleep in schizophrenics
and normals. Journal o f Nervous and M ental Diseases, 122, 263-269.
Dement, W ., Kahn, E., & Roffwarg, H . P. (1965). The influence of the laboratory
situation on the dreams o f the experimental subject. Journal o f Nervous and
M ental Disease, 140, 119-131.
REFERENCES 175
Dement, W ., & Kleitm an, N . (1957a). Cyclic variations in EEG during sleep and
their relation to eye movements, body m otility, and dreaming. Electroencepha
lography and C linical Neurophysiology , 9, 673-690.
Dement, W ., & Kleitman, N . (1957b). The relation of eye movements during sleep
to dream activity: A n objective method for the study of dreaming. Journal o f
Experim ental Psychology, 53, 339-346.
Dement, W ., & W olpert, E. (1958a). Relationships in the manifest content of
dreams occurring in the same night. Journal o f Nervous and M ental Disease,
126, 568-578.
Dement, W ., & W olpert, E. (1958b). The relation of eye movement, body m otility,
and external stim uli to dream content. Journal o f Experim ental Psychology,
44, 543-555.
Domhoff, G. W . (1969). Home dreams and laboratory dreams: Home dreams are
better. In M. Kramer (Ed.), Dream psychology and the ne<w biology o f dreaming
(pp. 119-217). Springfield, IL: C. C. Thomas.
Domhoff, G. W . (1993a). Personal dream histories: T heir theoretical and practical
relevance. Unpublished manuscript, University of California, Santa Cruz.
Domhoff, G. W . (1993b). The repetition of dreams and dream elements: A possible
clue to a function of dreams. In A . M offitt, M . Kramer, & R. Hoffmann
(Eds.), The functions o f dreams (pp. 293-320). Albany: State U niversity of
New York Press.
Domhoff, G. W . (1996). Finding meaning in dreams: A quantitative approach. New
York: Plenum Publishing.
Domhoff, G. W . (1999a). Drawing theoretical im plications from descriptive em piri
cal findings on dream content. Dreaming, 9, 201-210.
Domhoff, G. W . (1999b). New directions in the study of dream content using the
H a ll and Van de Castle coding system. Dreaming, 9, 115-137.
Domhoff, G. W . (1999c, June). Using H all/V an De Castle dream content analysis to
test new theories: A n example using a theory proposed by Ernest H artm ann. Paper
presented at the annual meeting of the Association for the Study of Dreams,
Santa Cruz, CA.
Domhoff, G. W ., & Gerson, A . (1967). Replication and critique of three studies
on personality correlates of dream recall. Journal o f Consulting Psychology,
31, 431.
Domhoff, G. W ., & Kamiya, J. (1964a). Problems in dream content study w ith
objective indicators: I. A comparison of home and laboratory dream reports.
Archives o f General Psychiatry, 11, 519-524-
Domhoff, G. W ., & Kamiya, J. (1964b). Problems in dream content study w ith
objective indicators: III. Changes in dream content throughout the night.
Archives o f General Psychiatry, 11, 529-532.
Domhoff, G. W ., & Schneider, A . (1998). New rationales and methods for quantita
tive dream research outside the laboratory. Sleep, 21, 398-404.
Domhoff, G. W ., & Schneider, A . (1999). Much ado about very little : The small
effect sizes when home and laboratory collected dreams are compared. Dream -
ing, 9, 139-151.
Doricchi, F., & Violani, C. (2000). Mesolimbic dopamine and the neuropsychology
o f dreaming: Some caution and reconsiderations. Behavioral and B rain Sciences,
23, 930-931.
Downing, J., & Marmorstein, R. (1973). Dreams and nightmares: A book o f G estalt
therapy sessions. New York: Harper & Row.
Dudley, L., & Fungaroli, J. (1987). The dreams o f students in a women’s college:
Are they different? ASD Newsletter, 4, 6-7.
Dudley, L., & Swank, M. (1990). A comparison of the dreams of college women
in 1950 and 1990. A S D Newsletter, 7, 3.
Dunlap, W ., Cortina, J., Vaslow, J., & Burke, M . (1996). Meta-analysis of experi
ments w ith matched groups or repeated measures designs. Psychological Methods,
1(2), 170-177.
Dunn, S. (2000). Dreams o f Derek: A H a ll/V a n de Castle coding and analysis. Unpub
lished manuscript, University of California, Santa Cruz.
Ekman, P. (1992a). A n argument for basic emotions. Cognition and Em otion, 6 (3 -
4), 169-200.
Ekman, P. (1992b). Are there basic emotions? Psychological Review, 99, 550-553.
Ellis, H. (1928). The synthesis of dreams: A study o f a series o f one hundred
dreams. In H. Ellis (Ed.), Studies in the psychology o f sex (Vol. 7, pp. 237-346).
Philadelphia: F. A . Davis Company.
Farley, F. H., Schmuller, J., & Fischbach, T. J. (1971). Dream recall and individual
differences. Perceptual and M otor Skills, 33, 379-384.
Fauconnier, G. (1997). Mappings in thought and language. New York: Cambridge
University Press.
Feinberg, I. (2000). REM sleep: Desperately seeking isomorphism. Behavioral and
Brain Sciences, 23, 931-934.
Ferguson, G. A . (1981). Statistical analysis in psychology and education. New York:
M cG raw-Hill.
Fisher, C. (1954). Dreams and perception: The role of preconscious and primary
modes o f perception in dream formation. Journal o f the Am erican Psychoanalytic
Association, 2, 389-445.
Fisher, S., & Greenberg, R. (1977). The scientific credibility o f Freud’s theories and
therapy. New York: Basic Books.
Fisher, S., & Greenberg, R. (1996). Freud scientifically appraised. New York: John
W iley.
Fiss, H. (1983). Toward a clinically relevant experimental psychology o f dreaming.
Hillside Journal o f C linical Psychiatry, 5, 147-159.
Fiss, H. (1986). A n empirical foundation for a self-psychology of dreaming. Journal
o f M in d and Behavior, 7(2-3), 161-191.
REFERENCES 177
Fiss, H . (1991). Experimental strategies for the study of the function o f dreaming.
In J. Antrobus & S. Ellman (Eds.), The m ind in sleep: Psychology and psychophysi
ology (2nd ed., pp. 308-326). New York: John W iley & Sons.
Fiss, H. (1993). The “ royal road” to the unconscious revisited: A signal detection
model o f dream function. In A . M offitt, M. Kramer, & R. Hoffm ann (Eds.),
The functions o f dreaming (pp. 381-418). Albany: State U niversity o f New
York Press.
Fiss, H., Ellman, S. J., & Klein, G. S. (1969). W aking fantasies following interrupted
and completed REM periods. Archives o f General Psychiatry, 21, 230-239.
Fiss, H., Klein, G. S., & Bokert, E. (1966). W aking fantasies following interruption
of two types o f sleep. Archives o f General Psychiatry, 14, 543-551.
Fitch, T., & Armitage, R. (1989). Variations in cognitive style among high and low
frequency dream recallers. Personality and Individual Differences, 10, 869-875.
Flanagan, O. (1995). Deconstructing dreams: The spandrels o f sleep. Journal o f
Philosophy, 92, 5-27.
Flanagan, O. (2000a). Dreaming is not an adaptation. Behavioral and B rain Sciences,
23, 936-939.
Flanagan, O. (2000b). Dream ing souls: Sleep, dreams, and the evolution o f the conscious
mind. New York: O xford University Press.
Fosse, R., Stickgold, R., & Hobson, J. (2001). Brain-m ind states: Reciprocal varia
tio n in thoughts and hallucinations. Psychological Science, 12, 30-36.
Foulkes, D. (1962). Dream reports from different states of sleep. Journal o f Abnorm al
and Social Psychology, 65, 14-25.
Foulkes, D. (1966). The psychology o f sleep. New York: Charles Scribner’s Sons.
Foulkes, D. (1967). Dreams of the male child: Four case studies. Journal o f C hild
Psychology and Psychiatry, 8, 81-98.
Foulkes, D. (1978). A grammar o f dreams. New York: Basic Books.
Foulkes, D. (1979). Home and laboratory dreams: Four em pirical studies and a
conceptual reevaluation. Sleep, 2, 233-251.
Foulkes, D. (1982). C hildren’s dreams. New York: W iley.
Foulkes, D. (1983). Cognitive processes during sleep: Evolutionary aspects. In A .
Mayes (Ed.), Sleep mechanisms and functions in humans and anim als: A n evolution
ary perspective (pp. 313-337). W okington, England: Van Nostrand Reinhold.
Foulkes, D. (1985). Dream ing: A cognitive-psychological analysis. Hillsdale, NJ: Law
rence Erlbaum.
Foulkes, D. (1990a). Dreaming and consciousness. European Journal o f Cognitive
Psychology, 2, 3 9 -5 5 .
Foulkes, D. (1990b). Reflective consciousness and dreaming. Contem porary Psychol
ogy, 35, 120-122.
Foulkes, D. (1993). Data constraints on theorizing about dream function. In A .
M offitt, M . Kramer, & R. Hoffmann (Eds.), The functions o f dreaming (pp.
11-20). Albany: State University o f New York Press.
Foulkes, D. (1996a). Dream research: 1953-1993. Sleep, 19, 609-624.
Foulkes, D. (1996b). Misrepresentation of sleep-laboratory dream research w ith
children. Perceptual and M otor Skills, 83, 205-206.
Foulkes, D. (1999). C hildren’s dreaming and the development o f consciousness. Cam
bridge, M A: Harvard University Press.
Foulkes, D., &. Fleisher, S. (1975). M ental activity in relaxed wakefulness. Journal
o f Abnorm al Psychology, 84, 66-75.
Foulkes, D., H ollifield, M., Sullivan, B., Bradley, L., & Terry, R. (1990). REM
dreaming and cognitive skills at ages 5-8: A cross-sectional study. International
Journal o f Behavioral Development, 13, 447-465.
Foulkes, D., Larson, J., Swanson, E., & Rardin, M. (1969). Two studies o f childhood
dreaming. Am erican Journal o f O rthopsychiatry, 39, 627-643.
Foulkes, D., Meier, B., Strauch, I., Kerr, N ., Bradley, L., & H ollifield, M. (1993).
Linguistic phenomena and language selection in the REM dreams of German-
English bilinguals. International Journal o f Psychology, 2 8 (6 ), 871-891.
Foulkes, D., Pivik, R., Steadman, H., Spear, P., & Symonds, J. (1967). Dreams of
the male child: A n EEG study. Journal o f Abnorm al Psychology, 72, 457-467.
Foulkes, D., & Rechtschaffen, A . (1964). Presleep determinants of dream content:
Effects of two films. Perceptual and M otor Skills, 19, 983-1005.
Foulkes, D., & Schmidt, M. (1983). Temporal sequence and u n it comparison
composition in dream reports from different stages o f sleep. Sleep, 6, 265-280.
Foulkes, D., & Scott, E. (1973). A n above-zero baseline for the incidence of
momentarily hallucinatory mentation. Sleep Research, 2, 108.
Foulkes, D., & Shepherd, J. (1971). M anual fo r the scoring system fo r children’s dreams
[Mimeograph]. Laramie: University of Wyoming.
Foulkes, D., Sullivan, B., Kerr, N ., & Brown, L. (1988). Appropriateness of dream
feelings to dreamed situations. Cognition and Em otion, 2, 29-39.
Foulkes, D., & Vogel, G. (1965). M ental activity at sleep onset. Journal o f Abnorm al
Psychology, 70, 231-243.
Frank, J. (1946). C linical survey and results o f 200 cases of prefrontal leucotomy.
Journal o f M ental Sciences, 92, 497-508.
Frank, J. (1950). Some aspects o f lobotomy (prefrontal leucotomy) under psychoana
lytic scrutiny. Psychiatry, 13, 35-42.
Franklin, R. D., A llison, D. B., & Gorman, B. S. (1997). Design and analysis o f
single-case research. Mahwah, NJ: Lawrence Erlbaum Associates.
Freud, S. (1900). The interpretation o f dreams (J. Crick, Trans.). London: Oxford
University Press.
Freud, S. (1912/1961). Totem and taboo. In J. Strachey (Ed. & Trans.), The standard
edition o f the complete psychological works o f Sigmund Freud (Vol. 13, pp. 1-161).
London: Hogarth Press. (O riginal work published 1912)
Freud, S. (1920/1961). Beyond the pleasure principle. In J. Strachey (Ed. &. Trans.),
The standard edition o f the complete psychology works o f Sigmund Freud (Vol. 18,
pp. 7-64). London: Hogarth Press. (O riginal work published 1920)
179
Freud, S. (1933/1961). New introductory lectures on psychoanalysis. In J. Strachey
(Ed. & Trans.), The standard edition o f the complete psychological works o f Sigmund
Freud (Vol. 22, pp. 7-30). London: Hogarth Press. (O riginal work published
in 1933)
Fromm, E. (1949). The nature of dreams. Scientific Am erican, 181(5), 44-47.
Fromm, E. (1951). The forgotten language. New York: Grove Press.
Fudin, R. (1999). Sublim inal psychodynamic activation: Methodological problems
and questions in Silverman’s experiments. Perceptual and M otor S kills, 89, 235-
244.
Gackenbach, J. (1988). The psychological content o f lucid dreams. In J. Gackenbach
<Sl S. LaBerge (Eds.), Conscious m ind, sleeping brain: Perspectives on lucid dreaming
(pp. 181-200). New York: Plenum.
Gackenbach, J., & Bosveld, J. (1989). Control your dreams. New York: Harper
& Row.
Giambra, L., Jung, R., & Grodsky, A . (1996). Age changes in dream recall in
adulthood. Dream ing, 6, 17-31.
Gibbs, R. (1994). The poetics o f m ind: Figurative thought, language, and understanding.
New York: Cambridge University Press.
Gibbs, R. (1999). Speaking and thinking w ith metonymy. In K. Panther & G.
Radden (Eds.), Metonymy in language and thought (pp. 61-75). Philadelphia:
John Benjamins.
G ironell, A ., Calzada, M ., Sagales, T., & Barraquer-Bordas, L. (1995). Absence of
REM sleep and altered non-REM sleep caused by a haematoma in the pontine
tegmentum. Journal o f Neurology, Neurosurgery, and Psychiatry, 59, 195-196.
Goodenough, D. (1991). Dream recall: History and current status of the field. In
S. Ellman & J. Antrobus (Eds.), The m ind in sleep: Psychology and psychophysiol'
ogy (2nd ed., pp. 143-171). New York: John W iley & Sons.
Gottesmann, C. (1999). Neurophysiological support of consciousness during waking
and sleep. Progress in Neurobiology, 59, 469-508.
Gottesmann, C. (2000). Each distinct type of mental state is supported by specific
brain functions. Behavioral and Brain Sciences, 23 , 941-943.
Gottschalk, L., & Gleser, G. (1969). The measurement o f psychological states through
the content analysis o f verbal behavior. Berkeley: U niversity of C alifornia Press.
Grady, J. (1999). A typology o f m otivation for conceptual metaphor: C orrelation
vs. resemblance. In R. Gibbs & G. Steen (Eds.), M etaphor in cognitive linguistics
(pp. 79-100). Philadelphia: John Benjamins.
Grady, J., Oakley, T ., & Coulson, S. (1999). Blending and metaphor. In R. Gibbs
& G. Steen (Eds.), M etaphor in cognitive linguistics (pp. 101-124). Philadelphia:
John Benjamins.
Greenberg, R., Katz, H., Schwartz, W ., & Pearlman, C. (1992). A research based
reconsideration of the psychoanalytic theory of dreaming. Journal o f the A m e ri'
can Psychoanalytic Association, 40, 531-550.
Greenberg, R., & Pearlman, C. (1993). A n integrated approach to dream theory:
Contributions from sleep research and clinical practice. In A . M offitt, M.
Kramer, & R. Hoffmann (Eds.), The functions o f dreaming (pp. 363-380).
Albany: State University of New York Press.
Greenberg, R., & Pearlman, C. (1999). The interpretation o f dreams: A classic
revisited. Psychoanalytic Dialogues, 9, 749-765.
Greenwald, A . G. (1992). New look 3: Unconscious cognition reclaimed. Am erican
Psychologist, 47, 766-119.
Greenwald, A . G., Draine, S. C., & Abrams, R. L. (1996). Three cognitive markers
of unconscious semantic activation. Science, 273, 1699-1702.
Gregor, T. (1981). A content analysis o f Mehinaku dreams. Ethos, 9, 353-390.
G riffith, R., Miyago, O., & Tago, A . (1958). The universality o f typical dreams:
Japanese vs. Americans. Am erican Anthropologist, 60, 1173-1179.
H all, C. (1947). Diagnosing personality by the analysis o f dreams. Journal o f A b n o r'
m al and Social Psychology, 42, 68-79.
H all, C. (1951). W hat people dream about. Scientific Am erican, 184, 60-63.
H all, C. (1953 a). A cognitive theory o f dream symbols. Journal o f General Psychology,
48, 169-186.
H all, C. (1953b). A cognitive theory of dreams. Journal o f General Psychology,
49, 273-282.
H all, C. (1953c). The meaning o f dreams. New York: M cG raw-Hill.
H all, C. (1956). Current trends in research on dreams. In D. Brower & L. A b t
(Eds.), Progress in clinical psychology (Vol. 2, pp. 239-257). New York: Grune
and Stratton.
H all, C. (1966a). A comparison of the dreams o f four groups o f hospitalized mental
patients w ith each other and w ith a normal population. Journal o f Nervous
and M ental Diseases, 143, 135-139.
H all, C. (1966b). Studies o f dreams collected in the laboratory and at home. Institute
o f Dream Research Monograph Series (No. 1). Santa Cruz, CA: Institute of
Dream Research.
H all, C. (1967). Caveat lector. Psychoanalytic Review, 54, 655-661.
H all, C. (1969a). Content analysis o f dreams: Categories, units, and norms. In G.
Gerbner (Ed.), The analysis o f communication content (pp. 147-158). New
York: W iley.
H all, C. (1969b). Normative dream content studies. In M. Kramer (Ed.), Dream
psychology and the new biology o f dreaming (pp. 175-184). Springfield, IL: Charles
C. Thomas.
H all, C. (1984). A ubiquitous sex difference in dreams, revisited. Journal o f Personal'
ity and Social Psychology, 46, 1109-1117.
H all, C., & Domhoff, G .W . (1963a). A ubiquitous sex difference in dreams. Journal
o f Abnorm al and Social Psychology, 66, 278-280.
REFERENCES 181
H all, C., & Domhoff, G. W . (1963b). Aggression in dreams. International Journal
o f Social Psychiatry, 9, 259-267.
H all, C., & Domhoff, G. W . (1964). Friendliness in dreams. Journal o f Social
H all, C., Domhoff, G. W ., Blick, K., & Weesner, K. (1982). The dreams o f college
men and women in 1950 and 1980: A comparison of dream contents and sex
differences. Sleep, 5, 188-194.
H all, C., & Lind, R. (1970). Dreams, life and literature: A study o f Franz Kafka.
Chapel H ill: U niversity o f N orth Carolina Press.
H all, C., & Nordby, V . (1972). The individual and his dreams. New York: New
Am erican Library.
H all, C., & Van de Castle, R. (1966). The content analysis o f dreams. New York:
Appleton-Century-Crofts.
Harlow, J., & Roll, S. (1992). Frequency of day residue in dreams of young adults.
Perceptual and M otor Skills, 74, 832-834.
Hartmann, E. (1968). The day residue: Time distribution of waking events. Psycho
physiology, 5, 222.
Hartmann, E. (1984). The nightmare. New York: Basic Books.
Hartmann, E. (1998). Dreams and nightmares. New York: Plenum.
Hartmann, E. (2000). The waking-tO'dreaming continuum and the effects of emo
tion. Behavioral and B rain Sciences, 23, 947-950.
Hartmann, E., Elkin, R., & Garg, M . (1991). Personality and dreaming: The dreams
o f people w ith very thick or very th in boundaries. Dream ing, 1, 311-324.
Hartmann, E., Rosen, R., & Rand, W . (1998). Personality and dreaming: Boundary
structure and dream content. Dreaming, 8, 31-39.
Hartmann, E., Russ, D., O ldfield, M ., Falke, R., & Skoff, B. (1980). Dream content:
Effects o f LDOPA. Sleep Research, 9, 153.
Hauri, P. (1975). Categorization of sleep mental activity for psychophysiological
studies. In G. Lairy & P. Salzarulo (Eds.), The experimental study o f sleep:
Methodological problems (pp. 271-281). New York: Elsevier Scientific Pub
lishing.
Heiss, W ., Pawlik, G., Herholz, K., Wagner, R., & Wienhard, K. (1985). Regional
cerebral glucose metabolism in man during wakefulness, sleep, and dreaming.
B rain Research, 327, 362-366.
Herman, J., Ellman, S., & Roffwarg, H. (1978). The problem o f NREM dream
recall reexamined. In A . A rkin , J. Antrobus, & S. Ellman (Eds.), The m ind in
sleep: Psychology and psychophysiology (pp. 59-62). Hillsdale, NJ: Erlbaum.
Heynick, F., & dejong, M. (1985). Dreams elicited by the telephone: A comparative
content analysis. In W . Koella, E. Ruther, & H. Schulz (Eds.), Sleep '84 (pp.
341-343). New York: Verlage.
H ill, C. E. (1996). W orking w ith dreams in psychotherapy. New York: G uilford Press.
REFERENCES
H ill, C. E., Kelley, F. A ., Davis, T. L., Crook, R. E., Maldonado, L. E., Turkson,
M. A ., et al. (2001). Predictors of outcome of dream interpretation sessions:
Volunteer client characteristics, dream characteristics, and type of interpreta
tion. Dreaming, 11(2), 53-72.
Hiscock, M ., & Cohen, D. (1973). Visual imagery and dream recall. Journal o f
Research in Personality, 72, 179-188.
Hobson, J. (1988). The dreaming brain. New York: Basic Books.
Hobson, J. (2000). The ghost of Sigmund Freud haunts M ark Solms’s dream theory.
Behavioral and B rain Sciences, 23, 951-952.
Hobson, J., Hoffman, S., Helfand, R., & Kostner, D. (1987). Dream bizarreness
and the activation-synthesis hypothesis. Hum an Neurobiology, 6, 157-164.
Hobson, J., & McCarley, R. (1977). The brain as a dream state generator: A n
activation-synthesis hypothesis of the dream process. Am erican Journal o f Psy
chiatry, 134, 1335-^1348.
Hobson, J., McCarley, R., & Wyzinski, P. (1975). Sleep cycle oscillation: Reciprocal
discharge by two brainstem neuronal groups. Science, 189, 55-58.
Hobson, J., Pace-Schott, E. F., & Stickgold, R. (2000a). Dream science 2000: A
response to commentaries on dreaming and the brain. Behavioral and Brain
Sciences, 23, 1019-1034.
Hobson, J., Pace-Schott, E. F., & Stickgold, R. (2000b). Dreaming and the brain:
Toward a cognitive neuroscience of conscious states. Behavioral and B rain
Sciences, 23, 793-842.
Hobson, J., Stickgold, R., &. Pace-Schott, E. F. (1998). The neuropsychology of
REM sleep dreaming. NeuroReport, 9, R1-R14.
Holland, D., & Kipnis, A . (1994). Metaphors for embarrassment and stories of
exposure: The not-so-egocentric self in American culture. Ethos, 22, 316-342.
Howard, M. (1978). M anifest dream content o f adolescents. Unpublished doctoral
dissertation, Iowa State University, Ames.
Howe, J. B., & Blick, K. (1983). Emotional content of dreams recalled by elderly
women. Perceptual and M otor Skills, 56, 31-34.
Howe, M. (2000). The fate o f early memories. Washington, DC: American Psychologi
cal Association.
H unt, H. (1989). The m ultiplicity o f dreams: M em ory, im agination, and consciousness.
New Haven, CT: Yale University Press.
H unt, H. (2000). New m ultiplicities of dreaming and REMing. Behavioral and B rain
Sciences, 23, 953-955.
H unt, H., O gilvie, R., Belicki, K., Belicki, D., & A talick, E. (1982). Forms of
dreaming. Perceptual and M otor Skills, 54, 559-633.
H unt, H., Ruzycki-Hunt, K., Pariak, D., & Belicki, K. (1993). The relationship
between dream bizarreness and imagination: A rtifa ct or essence? Dreaming,
3, 179-199.
Hunter, J. (1997). Needed: A ban on the significance test. Psychological Science,
8, 3-7.
REFERENCES 183
Hurovitz, C., Dunn, S., Domhoff, G. W ., & Fiss, H . (1999). The dreams o f blind
men and women: A replication and extension o f previous findings. Dream ing,
9, 183-193.
Janowsky, J., & Carper, R. (1996). Is there a neural basis for cognitive transitions
in schoohage children? In A . Sameroff & M. H a ith (Eds.), The five to seven
year shift (pp. 33-60). Chicago: University of Chicago Press.
Jouvet, M. (1999). The paradox o f sleep: The story o f dreaming. Cambridge, M A :
M IT Press.
Jung, C. (1963). Memories, dreams, reflections. New York: Pantheon.
Jung, C. (1974). Dreams. Princeton, NJ: Princeton University Press.
Jus, A ., Jus, K., Villenueve, A ., Pires, A ., Lachance, R., Fortier, J., et al. (1973).
Studies on dream recall in chronic schizophrenic patients after prefrontal
lobotomy. Biological Psychiatry, 6, 275-293.
Kamiya, J. (1961). Behavioral, subjective, and physiological aspects o f drowsiness
and sleep. In D. W . Fiske & S. R. Maddi (Eds.), Functions o f varied experience
(pp. 145-174). Homewood, IL: Dorsey.
Kane, C. M ., M ellen, R. R., Patten, P., & Samano, I. (1993). Differences in the
manifest dream content of Mexican, Mexican American, and Anglo Am erican
college women: A research note. Hispanic Journal o f Behavioral Sciences, 15,
134-139.
K err,N ., & Foulkes, D. (1981). Right hemispheric mediation of dream visualization:
A case study. C ortex, 17, 603-610.
Kerr, N., Foulkes, D., & Jurkovic, G. J. (1978). Reported absence o f visual dream
imagery in a norm ally sighted subject w ith Turner’s syndrome. Journal o f M ental
Imagery, 2, 247-264.
Ketchum, J., Sidell, F., Crowell, E., Aghajanian, G., & Hayes, A . (1973). A tropine,
scopolamine, and ditran: Comparative pharmacology and antagonists in man.
Psychopharmacologia, 28, 121-145.
Kirschner, N . (1999). M edication and dreams: Changes in dream content after
drug treatment. Dreaming, 9, 195-200.
Klinger, E., & Cox, W . (1987/1988). Dimensions of thought flow in everyday life.
Im agination, C ognition, and Personality, 7, 105-128.
Knudson, R. M ., & M inier, S. (1999). The on-going significance of significant
dreams: The case of the bodiless head. Dreaming, 9, 235-245.
Krakow, B., Kellner, R., Pathak, D., & Lambert, L. (1995). Imagery rehearsal
treatment for chronic nightmares. Behaviour Research and Therapy, 33, 837-
843.
Kramer, M. (1993). The selective mood regulatory function of dreaming: A n update
and revision. In A . M offitt, M. Kramer, & R. Hoffmann (Eds.), The functions
o f dreams (pp. 139-195). Albany: State University of New York Press.
Kramer, M. (1999, June). A review o f dreaming by psychiatric patients: A n update.
Paper presented at the annual meeting of the Association for the Study of
Dreams, Santa Cruz, CA.
Kramer, M. (2000a). Dreaming has content and meaning not just form. Behavioral
and Brain Sciences, 23, 959-961.
Kramer, M. (2000b). Dreams and psychopathology. In M. Kryger, T. Roth, & W .
Dement (Eds.), Principles and practices o f sleep medicine (3rd ed., pp. 511-519).
Philadelphia: W . B. Saunders.
Kramer, M., Kinney, L., & Scharf, M. (1983). Sex differences in dreams. Psychiatric
Journal o f the U niversity o f O ttaw a, 8, 1-4.
Kramer, M., & Roth, T. (1973). Comparison o f dream content in laboratory dream
reports of schizophrenic and depressive patient groups. Comprehensive Psychia
try, 14, 325-329.
Kramer, M., & Roth, T. (1979). Dreams in psychopathology. In B. W olman (Ed.),
Handbook o f dreams (pp. 361-367). New York: Van Nostrand Reinhold.
Kramer, M., Roth, T., &. Trinder, J. (1975). Dreams and dementia: A laboratory
exploration of dream recall and dream content in chronic brain syndrome
patients. International Journal o f Aging and Hum an Development, 6, 169-178.
Kramer, M., Schoen, L., & Kinney, L. (1987). Nightmares in Vietnam veterans.
Journal o f the Am erican Academy o f Psychoanalysis, 15, 67-81.
Krohn, A., & Mayman, M. (1974). Object relations in dreams and projective tests.
B ulletin o f the Menninger C linic, 38, 445-466.
Kuiken, D., & Sikora, S. (1993). The impact o f dreams on waking thoughts and
feelings. In A . M offitt, M. Kramer, & R. Hoffmann (Eds.), The functions o f
dreaming (pp. 419-476). Albany: State U niversity o f New York Press.
LaBerge, S. (1985). Lucid dreaming. Los Angeles: J. P. Tarcher.
LaBerge, S., Nagel, L., Dement, W ., & Zarcone, V. (1981). Lucid dreaming verified
by volitional communication during REM sleep. Perceptual and M o to r Skills,
52, 727-732.
Lakoff, G. (1987). W omen, fire , and dangerous things. Chicago: U niversity of C h i
cago Press.
Lakoff, G. (1993a). How metaphor structures dreams. Dreaming, 3, 77-98.
Lakoff, G. (1993b). The contemporary theory o f metaphor. In A . O rtony (Ed.),
Metaphor and thought (2nd ed., pp. 202-251). New York: Cambridge U niver
sity Press.
Lakoff, G. (1997). How unconscious metaphorical thought shapes dreams. In D.
Stein (Ed.), Cognitive science and the unconscious (pp. 89-120). Washington,
DC: Am erican Psychiatric Press.
Lakoff, G., & Johnson, M . (1999). Philosophy in the flesh. New York: Basic Books.
Lakoff, G., & Turner, M . (1989). More than cool reason. Chicago: U niversity of
Chicago Press.
Latta, C. (1998). The manifest dream content of premenarcheal and postmenarcheal
girls. Dissertation Abstracts International, 5 9 (6 ), 3064B.
Lavie, P. (1984). Localized pontine lesion: Nearly total absence of REM sleep.
Neurology, 34, 118-120.
REFERENCES 185
Lavie, P. (1990). Penile erections in a patient w ith nearly total absence o f REM:
A follow-up study. Sleep, 13, 276-278.
Leach, C. (1979). Introduction to statistics. New York: W iley.
LeDoux, J. (1996). The em otional brain. New York: Simon & Schuster.
Levin, R. (2000). Nightmares: Friend or foe? Behavioral and B rain Sciences, 23, 965.
Lilienfeld, S., Wood, J., & Garb, H. (2000). The scientific status of projective
techniques. Psychological Science in the Public Interest, 1(2), 27-66.
Lindorff, D. (1995). One thousand dreams: The spiritual awakening of Wolfgang
Pauli. Journal o f A n a lytica l Psychology, 40, 555-569.
Livingston, G., & Levin, R. (1991). The effects o f dream length on the relationship
between primary process in dreams and creativity. Dreaming, 1, 301-309.
Llinas, R., & Pare, D. (1991). O f dreaming and wakefulness. Neuroscience, 44, 521-
535.
Loftus, E., Joslyn, S., & Polage, D. (1998). Repression: A mistaken impression?
Development and Psychopathology, 10, 781-792.
Loftus, E., & Ketcham, K. (1994). The m yth o f repressed memory. New York: St.
M artin’s Press.
Loftus, E., & Polage, D. (1999). Repressed memories: When are they real? How
are they false? Psychiatric C linics o f N orth Am erica, 22, 61-70.
Lortie-Lussier, M ., Cote, L., & Vachon, J. (2000). The consistency and continuity
hypotheses revisited through the dreams o f women at two periods o f their
lives. Dream ing, 10, 6 1 -1 6 .
Lortie-Lussier, M ., Schwab, C., & de Koninck, J. (1985). W orking mothers versus
homemakers: Do dreams reflect the changing roles of women? Sex Roles, 12,
1009-1021.
Lortie-Lussier, M ., Simond, S., Rinfret, N., & De Koninck, J. (1992). Beyond sex
differences: Family and occupational roles’ impact on women’s and men’s
dreams. Sex Roles, 26, 79-96.
Luria, A . (1973). The w orking brain: An introduction to neuropsychology. New
York: Penguin.
Maharaj, N . (1997). An investigation into the content and structure o f schizophrenic
dreams. Unpublished doctoral dissertation, Leiden University, Leiden, The
Netherlands.
Mahowald, M ., & Schenck, C. (2000). REM sleep parasomnias. In M. Kryger, T.
Roth, & W . Dement (Eds.), Principles and practices o f sleep medicine (3rd ed.,
pp. 724-741). Philadelphia: W . B. Saunders.
Mamelak, A . N ., & Hobson, J. (1989). Dream bizarreness as the cognitive correlate
o f altered neuronal behavior in REM sleep. Journal o f Cognitive Neuroscience,
1 , 201- 222.
Maquet, P. (2000). Functional neuroimaging o f normal human sleep by positron
emission tomography. Journal o f Sleep Research, 9, 207-231.
Maquet, P., Peters, J. M., Aerts, J., Delfiore, G., Dequerldre, C., Luxen, A ., et al.
(1996). Functional neuroanatomy o f human rapid-eye-movement sleep and
dreaming. N ature, 383, 163-166.
Markman, A . B. (1999). Knowledge representation. Mahwah, NJ: Erlbaum.
Marquardt, C. J. G., Bonato, R. A ., & Hoffmann, R. (1996). A n empirical investiga
tion into the day-residue and dream-lag effects. Dreaming, 6, 57-65.
M artinelli, R. (1983). Dream recall, imaginal processes, and short-term memory:
A p ilo t study. Perceptual and M otor Skills, 57, 718.
M athur, R., & Douglas, N . (1995). Frequency o f EEG arousals from nocturnal sleep
in normal subjects. Sleep, 18, 330-333.
M atlock, T. (1988, August). The metaphorical extension o f “see.” Paper presented at
the proceedings o f the Western Conference on Linguistics, Berkeley, C A.
M atlock, T., & Sweetser, E. (1989). Semantic change o f perception verbs to evidentials
and m ental state verbs. Unpublished manuscript, University of California,
Santa Cruz.
M attoon, M. (1978). Applied dream analysis: A Jungian approach. New York: W iley
& Sons.
Mazziotta, J. C. E., Toga, A . W . E., & Frackowiak, R. S. J. E. (Eds.). (2000). Brain
mapping: The disorders. San Diego, CA: Academic Press.
Mazzoni, G., &. Loftus, E. (1998). Dreaming, believing, and remembering. In J. de
Rivera & T. Sarbin (Eds.), Believed-in imaginings: The narrative construction o f
reality (pp. 145-156). Washington, DC: American Psychological Association.
Mazzoni, G., Loftus, E., Seitz, A ., & Lynn, S. (1999). Changing beliefs and memories
through dream interpretation. Applied Cognitive Psychology, 13, 125-144.
McCarley, R., & Hobson, J. (1975). Discharge patterns o f cat pontine brain stem
neurons during desynchronized sleep. Journal o f Neurophysiology, 38, 751-766.
Meehl, P. (1997). The problem is epistemology, not statistics: Replace significance
tests by confidence intervals and quantify accuracy of risky numerical predic
tions. In L. Harlow, S. M uliak, & J. Steiger (Eds.), W hat if there were no
significance tests? (pp. 392-425). Mahwah, NJ: Erlbaum Associates.
Meier, B. (1993). Speech and thinking in dreams. In C. Cavallero & D. Foulkes
(Eds.), Dreaming as cognition (pp. 58-76). New York: Harvester Wheatsheaf.
Meier, C. (Ed.). (1992). Wolgang Pauli und C. G. Jung: Ein Briefwechsel 1932-1958
[Letters between Pauli and Jung, 1932-1958]. Berlin: Springer.
Meier, C., Ruef, H., Zeigler, A ., & H all, C. (1968). Forgetting of dreams in the
laboratory. Perceptual and M o to r Skills, 26, 551-557.
M erritt, J., Stickgold, R., Pace-Schott, E. F., W illiam s, J., & Hobson, J. (1994).
Emotion profiles in the dreams of men and women. Consciousness and Cognition,
3, 46-60.
M icceri, T. (1989). The unicorn, the normal curve, and other improbable creatures.
Psychological B ulletin, 105, 156-166.
Mooney, C., & Duval, R. (1993). Bootstrapping: A nonparametric approach to statistical
inference. Newbury Park, C A : Sage Publications.
REFERENCES 187
Morrison, A . R., & Sanford, L. D. (2000). C ritica l brain characteristics to consider in
developing dream and memory theories. Behavioral and B rain Sciences, 23, 977.
Murphy, G., & Lassaline, M. (1997). Hierarchical structure in concepts and the
basic level o f categorization. In K. Lamberts & D. Shanks (Eds.), Knowledge,
concepts, and categories (pp. 93-131). Cambridge, M A : M IT Press.
M urri, L., Massetani, R., Siciliano, G., & Arena, R. (1985). Dream recall after
sleep interruption in brain-injured patients. Sleep, 8, 356-362.
Nanna, M., & Sawilowsky, S. (1998). Analysis of Likert scale data in disability
and medical rehabilitation research. Psychological Methods, 3, 55-67.
Neher, A . (1996). Jung’s theory o f archetypes: A critique. Journal o f Hum anistic
Nielsen, T., & Powell, R. (1992). The day-residue and dream-lag effect. Dreaming,
2, 67-77.
Nielsen, T., Zadra, A., Germain, A ., & M ontplaisir, J. (1999). The typical dreams
o f sleep patients: Consistent profile w ith 284 new cases. Sleep, 22, S177.
Nielsen, T. (2000a). A review o f mentation in REM and NREM sleep: “Covert”
REM sleep as a possible reconciliation o f two opposing models. Behavioral and
B rain Sciences, 23, 851-866.
Nielsen, T. (2000b). Covert REM sleep effects on REM mentation: Further method
ological considerations and supporting evidence. Behavioral and Brain Sciences,
23, 1040-1057.
Nielsen, T., & Germain, A . (2000). Post-traumatic nightmares as a dysfunctional
state. Behavioral and B rain Sciences, 23, 978-979.
Nishigawa, N ., Brubaker, L., & Domhoff, G. W . (2001). The dreams o f Japanese
women: A replication and extension. Unpublished manuscript, University of
C alifornia, Santa Cruz.
Nofzinger, E., Berman, S., Fasiczka, A ., Miewald, J. M ., Meltzer, C. C., Price, J.
C., et al. (2001). Effects o f bupropion SR on anterior paralimbic function
during waking and REM sleep in depression: Preliminary findings using (18F)-
FDG PET. Psychiatry Research, 106, 95-111.
Nofzinger, E., M intun, M ., Wiseman, M ., Kupfer, D., & Moore, R. (1997). Forebrain
activation in REM sleep: A n FDG PET study. B rain Research, 770, 192-201.
Nofzinger, E., Nichols, T. E., Meltzer, C. C., Price, J., Steppe, D. A ., Miewald, J.
M ., et al. (1999). Changes in forebrain function from waking to REM sleep
in depression: Preliminary analyses o f (18F)-FDG PET studies. Psychiatry Re
search, 91, 59-78.
Nomura, M. (1996). The ubiquity o f the fluid metaphor in Japanese: A case study.
Poetica, 46, 41-75.
Noreen, E. (1989). Computer intensive methods fo r testing hypotheses: A n introduction.
New York: John W iley & Sons.
Ofshe, R., & Watters, E. (1994). M aking monsters: False memories, psychotherapy,
and sexual hysteria. New York: Charles Scribner’s Sons.
O gilvie, R. (1982). Lucid dreaming and alpha a ctivity: A prelim inary report. Percept
tual and M otor Skills, 55(3, Pt. 1), 795-808.
O gilvie, R., H unt, H ., Sawicki, C., & Samahalskyi, J. (1982). Psychological corre-
lates o f spontaneous m iddle ear muscle a c tiv ity during sleep. Sleep, 5, 11-27.
O gilvie, R., Takeuchi, T ., & M urphy, T . (2000). Expanding Nielsen’s covert REM
model, questioning Solms’s approach to dreaming and REM sleep, and re in ter
preting the Vertes and Eastman view o f REM sleep and memory. Behavioral
and Brain Sciences, 23, 981-983.
O ’N e ll, C., & O ’N e ll, N . (1977). A cross-cultural comparison o f aggression in
dreams: Zapotecs and Am ericans. International Journal o f Social Psychiatry,
125, 35-41.
Osgood, C. (1959). The representation model and relevant research methods. In
I. de Sola Pool (Ed.), Trends in content analysis (pp. 54-78). Urbana: U nive rsity
o f Illin o is Press.
Osgood, C., May, W ., &. M iron , M . (1975). Cross-cultural universals o f affective
meaning. Urbana: U nive rsity o f Illin o is Press.
Pace-Schott, E. F. (2000). N ielsen’s concept o f covert REM sleep is a path toward
a more realistic view o f sleep psychophysiology. Behavioral and B rain Sciences,
23, 983-984.
Paolino, A . F. (1964). Dreams: Sex differences in aggressive content. Journal o f
Projective Techniques and Personality Assessment, 28, 219-226.
Paus, T ., Zijdenbos, A ., W orsely, K ., C ollins, D., Blum enthal, J., Giedd, J., et al.
(1999). Structural m aturation o f neural pathways in children and adolescents:
In vivo study. Science, 283, 1908-1911.
Pennebaker, J. W ., & Graybeal, A . (2001). Patterns o f natural language use: D isclo
sure, personality, and social integration. Current Directions in Psychological
Science, 10, 90-93.
Pennebaker, J. W ., &. Keough, K. A . (1999). Revealing, organizing, and reorganizing
the self in response to stress and em otion. In R. Contrada & R. Ashmore
(Eds.), Self, social identity, and physical health: Interdisciplinary explorations, (pp.
101-121). New York: O xford U niversity Press.
Pennebaker, J. W ., & Seagal, J. D. (1999). Forming a story: The health benefits
o f narrative. Journal o f C linical Psychology, 55, 1243-1254.
Perry, E., & Piggott, M . (2000). N eurotransm itter mechanisms o f dreaming: Im plica
tio n o f modulatory systems based on dream intensity. Behavioral and Brain
Sciences, 23, 990-992.
Perry, E., W alker, M ., Grace, J., & Perry, R. (1999). A cetylcholine in m ind: A
neurotransm itter correlate o f consciousness? Trends in Neurosciences, 22, 2 7 3 -
280.
P ivik, R. T . (1986). Sleep: Physiology and psychophysiology. In M . Coles, E.
D onchin, & S. Porges (Eds.), Psychophysiology: Systems, processes and applications
(pp. 378-406). New York: G u ilfo rd Press.
189
P ivik, R. T . (2000). Psychophysiology o f dreams. In M . Kryger, T . R oth, & W .
Dement (Eds.), Principles and practices o f sleep medicine (3rd ed., pp. 491-501).
Philadelphia: W . B. Saunders.
Popp, C., Luborsky, L., & C rits-C hristoph, P. (1992). The parallel o f the C C R T
from therapy narratives w ith the C C R T from dreams. In L. Luborsky & P.
C rits-C hristoph (Eds.), Understanding transference (pp. 158-172). New York:
Basic Books.
Rechtschaffen, A . (1978). The single-mindedness and isolation o f dreams. Sleep,
1, 97-109.
Rechtschaffen, A . (1997). Postscript, 1995: The single-mindedness and isolation
o f dreams. In M . Myslobodsky (Ed.), The mythomanias: The nature o f deception
and self'deception (pp. 219-223). Mahwah, NJ: Lawrence Erlbaum Associates.
Reichers, M ., Kramer, M ., & Trinder, J. (1970). A replication o f the H a ll-V a n de
Castle character scale norms. Psychophysiology, 7, 238.
Reinsel, R., A ntrobus, J., & W ollm an, M . (1992). Bizarreness in dreams and waking
fantasy. In J. A ntrobus & M . B e rtin i (Eds.), The neuropsychology o f sleep and
dreaming (pp. 157-184). H illsdale, NJ: Erlbaum.
Reinsel, R., W ollm an, M ., &. Antrobus, J. (1986). Effects o f environm ental context
and co rtica l activatio n on thought. Journal o f M in d and Behavior, 7, 259-275.
Reis, W . (1959). A comparison o f the interpretation o f dream series w ith and
w ith o u t free associations. In M . F. D eM artino (Ed.), Dreams and personality
dynamics (pp. 211-225). Springfield, IL : Charles C. Thomas.
Resnick, J., Stickgold, R., Rittenhouse, C., & Hobson, J. (1994). Self-representation
and bizarreness in ch ild re n ’s dream reports collected in the home setting.
Consciousness and C ognition, 3, 30-45.
Revonsuo, A . (2000). The reinterpretation o f dreams: A n evolutionary hypothesis
o f the fu n ctio n o f dreaming. Behavioral and B rain Sciences, 23, 877-901.
Revonsuo, A ., & S alm ivalli, C. (1995). A content analysis o f bizarre elements in
dreams. Dream ing, 5, 169-187.
Reynolds, H . (1984). Analysis o f nom inal data. Newbury Park, C A : Sage Publications.
Rittenhouse, C ., Stickgold, R., & Hobson, J. (1994). C onstraint on the transform a
tio n o f characters, objects, and settings in dream reports. Consciousness and
C ognition, 3, 100-113.
R ivkin , M . (2000). Developm ental neuroimaging o f children using m agnetic reso
nance techniques. M ental Retardation and Developmental Disabilities Research
Reviews, 6, 68-80.
Roberts, C. W . (1997). T ext analysis fo r the social sciences: Methods fo r drawing
statistical inferences from texts and transcripts. Mahwah, NJ: Erlbaum.
Robertson, L. (1998). Visuospatial attention and parietal function: T h e ir role in
object perception. In R. Parasuraman (Ed.), The attentive brain (pp. 257-278).
Cambridge, M A : M IT Press.
Rogoff, B. (1990). Apprenticeship in thinking: Cognitive development in social context.
New York: O xford U n ive rsity Press.
Rosenthal, R., Rosnow, R. L., & Rubin, D. B. (2000). Contrasts and effect sizes in
behavioral research: A correlational approach. New York: Cambridge U n ive r
sity Press.
Rosenthal, R., & R ubin, D. B. (1982). A simple general purpose display o f magnitude
o f experim ental effect. Journal o f Educational Psychology, 74, 166-169.
Rosnow, R. L., & Rosenthal, R. (1997). People studying people: A rtifacts and ethics
in behavioral research. New York: W . H . Freeman.
Roth, T ., Kramer, M ., &. Salis, P. (1979). Drugs, REM sleep, and dreams. In B.
W olm an (Ed.), Handbook o f dreams (pp. 203-225). New York: V an
Norstrand Reinhold.
Roussy, F. (2000). Testing the n o tio n o f co ntin u ity between waking experience
and REM dream content. Dissertation Abstracts International, 6 1 (2 ), 1106B.
Roussy, F., Brunette, M ., M ercier, P., G onthier, I., G renier, J., Sirois-Berliss, M .,
et al. (2000). D aily events and dream content: Unsuccessful m atching attempts.
Dreaming, 10, 77-83.
Roussy, F., Raymond, I., & De K oninck, J. (2000). A ffe c t in REM dreams: Explora
tio n o f a tim e -o f-n igh t effect. Sleep, 23, A 174-A 175.
Rubenstein, K ., & Krippner, S. (1991). Gender differences and geographical d iffe r
ences in content from dreams e licited by a television announcement. Interna
tional Journal o f Psychosomatics, 38, 40-44.
Salin-Pascual, R., Gerashchenko, D., & Shirom ani, P. J. (2000). Some myths are
slow to die. Behavioral and Brain Sciences, 23, 999-1000.
Saline,S. (1999). The most recent dreams o f children ages 8 -1 1. Dreaming, 9, 173-
181.
Sauvageau, A ., Nielsen, T., & M ontplaisir, J. (1998). Effects o f somatosensory
stim ulation on dream content in gymnasts and co ntro l participants: Evidence
o f vestibulom otor adaptation in REM sleep. Dreaming, 8, 125-134.
Sawilowsky, S. S., & B lair, R. C. (1992). A more realistic look at the robustness and
Type II error properties o f the t test to departures from population norm ality.
Psychological Bulletin, 111, 352-360.
Scarr, S. (1997). Rules o f evidence: A larger context for the statistical debate.
Psychological Science, 8, 16-17.
Schm idt, F. (1996). S tatistical significance testing and cum ulative knowledge in
psychology: Im plications for tra in ing o f researchers. Psychological Methods, 1,
115-129.
Schneider, A ., & Dom hoff, G. W . (1995, updated 2002). The quantitative study o f
dreams. Retrieved July 29, 2002, from http://www.dream research.net/
Schneider, A ., & Dom hoff, G . W . (1999, updated 2002). DreamBank. Retrieved

July 29, 2002, from http://www.dream bank.net/
Schredl, M . (2000). Dream research: Integration o f physiological and psychological
models. Behavioral and Brain Sciences, 23, 1001-1003.
REFERENCES 191
Schredl, M ., & Montasser, A . (1996). Dream recall: State or tra it variable? Part I:
M odel, theories, m ethodology and tra it factors. Im agination, C ognition and
Personality, 16, 181-210.
Schredl, M ., & Montasser, A . (1997). Dream recall: State or tra it variable? Part
II: State factors, investigations and fin a l conclusions. Im agination, C ognition
and Personality, 16 , 239-261.
Shapiro, D., W u, J., Hong, C ., Buchsbaum, M ., G ottschalk, L., Thom pson, V .,
H illya rd , D ., et al. (1995). Exploring the relationship between having co ntro l
and losing co ntro l to fu n ctio n a l neuroanatomy during the sleeping state. Psy-
chologia, 38, 133-145.
Shaver, P., Schwarz, J., Kirson, D., & O ’C onnor, D. (1987). Em otion knowledge:
Further explorations o f prototype approach. Journal o f Personality and Social
Psychology, 52, 1061-1086.
Sheppard, E. (1963). Systematic dream studies: C lin ic a l judgm ent and objective
measurements o f ego strength. Comprehensive Psychiatry, 4, 263-270.
Sheppard, E. (1969). Dream content analysis. In M . Kram er (Ed.), Dream psychology
and the new biology o f dreaming (pp. 225-254). S pringfield, IL: Charles C.
Thomas.
S hevrin, H . (1986). S ublim inal perception and dreaming. Journal o f M in d and
Behavior, 7 (2 -3 ), 379-395.
S hevrin, H . (1996). Conscious and unconscious processes: Psychodynamic, cognitive,
and neurophysiological convergences. New York: G u ilfo rd Press.
S hevrin, H ., & Eiser, A . (2000). C ontinued v ita lity o f the Freudian theory o f
dreaming. Behavioral and Brain Sciences, 23, 1004-1006.
S hevrin, H ., & Fisher, C. (1967). Changes in the effects o f a waking sublim inal
stimulus as a fu n ctio n o f dreaming and nondreaming sleep. Journal o f Abnorm al
Siegel, S., & Castellan, N . J. (1988). Nonparam etric statistics fo r the behavioral sciences
(2nd ed.). New York: M cG raw -H ill.
S lipp, S. (2000). S ublim inal stim ulation research and its im plications for psychoana
ly tic theory and treatm ent. Journal o f the Am erican Academy o f Psychoanalysis,
28, 305-320.
S m ith, C. (1995). Sleep states and memory processes. Behavioural B rain Research,
6 9 (1 -2 ), 137-145.
S m ith, C. (2000). C on te n t analysis and narrative analysis. In H . Reis & C. Judd
(Eds.), Handbook o f research methods in social and personality psychology (pp.
313-335). New York: Cambridge U niversity Press.
S m ith, M ., & H a ll, C. (1964). A n investigation o f regression in a long dream series.
Journal o f G erontology, 19, 66-71.
Snyder, F. (1970). T he phenomenology o f dreaming. In L. Madow & L. Snow
(Eds.), The psychodynamic implications o f the physiological studies on dreams (pp.
124-151). Springfield, IL : C.C. Thomas.
Solms, M . (1997). The neuropsychology o f dreams: A clinico-anatom ical study. Mahwah,
NJ: Lawrence Erlbaum.
Solms, M . (2000). Dreaming and REM sleep are controlled by different brain
mechanisms. Behavioral and Brain Sciences, 23, 843-850.
States, B. (1987). The rhetoric o f dreams. Ithaca, N Y : C ornell U nive rsity Press.
Stekel, W . (1911). Die sprache des traumes [The language o f dreams]. Wiesbaden,
Germany: J. F. Bergmann.
Steriade, M . (2000). B rain electrical a ctivity and sensory processing during waking
and sleep states. In M . Kryger, T . Roth, & W . Dement (Eds.), Principles and
practices o f sleep medicine (3rd ed., pp. 93-111). Philadelphia: W . B. Saunders.
Stewart, L., Ellison, A ., W alsh, V ., & Cowey, A . (2001). The role o f transcranial
magnetic stim ulation (T M S ) in studies o f vision, a tte n tio n and cognition.
A cta Psychologica, 1 0 7 (1 -3 ), 275-291.
S tickel, E. (1956). Dream frequency and personality variables. Unpublished doctoral
dissertation, Case W estern Reserve U niversity, Cleveland, OH.
Stickgold, R., Pace-Schott, E. F., & Hobson, J. (1994). A new paradigm for dream
research: M entation reports follow ing spontaneous arousal from REM and
N REM sleep recorded in a home setting. Consciousness and Cognition: A n
International Journal, 3, 16-29.
S tickgold, R., Scott, L., Rittenhouse, C., & Hobson, J. (1999). Sleep-induced
changes in associative memory. Journal o f Cognitive Neuroscience, 11 , 182-193.
Strauch, I. (1969, August). Psychological aspects o f dream recall. Paper presented at
the 19th International Congress o f Psychology, London.
Strauch, I. (1996, June). Self-representation in R E M dreams and waking fantasies at
ages 9 -1 1 and 1 1 -1 3 : A longitudinal study. Paper presented at the annual
meeting o f the Association fo r the Study o f Dreams, Berkeley, C A .
Strauch, I., & Lederbogen, S. (1999). The home dreams and waking fantasies o f
boys and girls ages 9 -1 5. Dreaming, 9, 153-161.
Strauch, I., & M eier, B. (1996). In search o f dreams: Results o f experimental dream
research. A lbany: State U niversity o f New Y ork Press.
Sweetser, E. (1990). From etymology to pragmatics. New York: Cambridge U n iv e r
sity Press.
Takeuchi, T ., O gilvie, R., Ferrelli, A . V ., M urphy, T . I., & B elicki, K. (2001).
The Dream Property Scale: A n exploratory English version. Consciousness and
Cognition: A n International Journal, 10, 341-355.
Tedlock, B. (1991). The new anthropology o f dreaming. Dreaming, 1, 161-178.
Thatcher, R. (Ed.). (1996). Developmental neuroimaging. San Diego, C A : A ca
demic Press.
Thom pson, B. (1999a). I f statistical significance tests are broken/misused, w hat
practices should supplement or replace them? Theory and Psychology, 9, 165 -
181.
Thom pson, B. (1999b). S tatistical significance tests, effect size reporting and the
vain pursuit o f pseudo-objectivity. Theory and Psychology, 9, 191-196.
REFERENCES 193
Thompson, B. (1999c). W hy “ encouraging” effect size reporting is n o t working:
The etiology o f researcher resistance to changing practices. Journal o f Psychol'
ogy, 133, 133-140.
Thompson, N . (2000). E volutionary psychology can ill afford adaptionist and men-
ta list credulity. Behavioral and Brain Sciences, 23, 1013-1014.
Tonay, V . (1990/1991). C alifo rn ia women and th e ir dreams: A historical and sub
cu ltural comparison o f dream content. Im agination, C ognition, and Personality,
10, 83-97.
Tonay, V . (1993). Personality correlates o f dream recall: W ho remembers ?Dreaming,

3, 1-8.
Torda, C. (1969). Dreams o f subjects w ith loss o f memory for recent events. Psycho'
physiology, 6, 358-365.
T rinder, J., & Kramer, M . (1971). Dream recall. Am erican Journal o f Psychiatry,
128, 296-301.
Trosman, H ., Rechtschaffen, A ., O ffenkrantz, W ., & W olpert, E. (1960). Studies

in the psychophysiology o f dreams. IV : Relations among dreams in sequence.
Archives o f General Psychiatry, 3, 602-607.
T ru p in , E. (1976). Correlates o f ego-level and agency-communion in stage REM

dreams o f 11-13 year old children. Journal o f C hild Psychology and C hild Psychia'
try , 17, 169-180.
Tyson, P. D., O gilvie , R., & .H u n t, H . (1984). Lucid, prelucid, and non lu cid dreams
related to the amount o f EEG alpha a c tiv ity during REM sleep. Psychophysiol'
ogy, 21, 442-451.
U rbina, S. P. (1981). M ethodological issues in the quantitative analysis o f dream

content. Journal o f Personality Assessment, 45, 71-78.
V an de Castle, R. (1969). Problems in applying methodology o f content analysis,
in M . Kramer (Ed.), Dream psychology and the new biology o f dreaming (pp.
185-197). Springfield, IL: Charles C. Thomas.
Van de Castle, R. (1983). A n im a l figures in fantasies and dreams. In A . Katcher
& A . Beck (Eds.), N ew perspectives on our lives ivith companion animals (pp.
148-173). Philadelphia: U niversity o f Pennsylvania Press.
V an d er K olk, B., B litz, R., Burr, W ., Sherry, S., & Hartm ann, E. (1984). Nightm ares
and trauma: A comparison o f nightmares after combat w ith life lo ng nightmares
in veterans. Am erican Journal o f Psychiatry, 141(2), 187-190.
Van Rompay, T . (2000). Emma’s dreams. Undergraduate senior thesis, Departm ent
o f Psychology, Leiden U niversity, Leiden, The Netherlands.
Vaughan, C. (1963). The development and use o f an operant technique to provide
evidence fo r visual imagery in the rhesus monkey under sensory deprivation. U npub
lished doctoral dissertation, U nive rsity o f Pittsburgh, Pittsburgh, PA.
Verdone, P. (1965). Tem poral reference o f m anifest dream content. Perceptual and
M o to r Skills, 20, 1253-1268.
REFERENCES
Vogel, G. (1991). Sleep-onset m entation; In S. Ellm an & J. Antrobus (Eds.), The
mind in sleep: Psychology and psychophysiology (2nd ed., pp. 125-136). New
York: W iley & Sons.
Vogel, G. (2000). C ritiqu e o f current dream theories. Behavioral and Brain Sciences,
23, 1014-1016.
Vogel, G ., Barrowclough, B., & Giesler, D. (1972). Lim ited discrim in ab ility of
REM and sleep onset reports and its psychiatric im plications. Archives o f General
Psychiatry, 26, 449-455.
von Meyenn, K. (Ed.). (1993). W . Pauli: Scientific correspondence w ith Bohr, Einstein,
Heisenberg 1940-1949. (V ol. 3). B erlin: Springer.
W ard, C., Beck, A ., &Rascoe, E. (1961). T ypical dreams: Incidence among psychiat
ric patients. Archives o f General Psychiatry, 5, 606-615.
W ebb, E., Campbell, D ., Schwartz, R., Sechrest, L., & Grove, J. (1981). Nonreactive
measures in the social sciences (2nd ed.). Chicago: Rand M cN ally.
Weed, S., & Hallam , F. (1896). A study o f dream-consciousness. Am erican Journal
o f Psychology, 7, 405-411.
W einstein, L., Schwartz, D., & A rk in , A . (1991). Q ualitative aspects o f sleep
m entation. In S. Ellm an & J. Antrobus (Eds.), The mind in sleep: Psychology
and psychophysiology (pp. 172-213). New York: John W iley & Sons.
Weisz, R., & Foulkes, D. (1970). Home and laboratory dreams collected under
uniform sampling conditions. Psychophysiology, 6, 588-596.
W elsh, M ., Pennington, B., &. Groisser, D. (1993). A norm ative-developm ental
study o f executive function: A w indow on prefrontal fun ctio n in children.
Developmental Neuropsychology, 7, 131-149.
W halen, P. (1998). Fear, vigilance, and am biguity: In itia l neuroimaging studies o f
the human amygdala. C urrent Directions in Psychological Science, 7, 177-188.
W hitm an, R., Kramer, M ., & Baldridge, B. (1963). W h ich dream does the patient
tell? Archives o f General Psychiatry, 8, 277-282.
W hitm an, R., Pierce, C., Maas, J., & Baldridge, B. (1961). Drugs and dreams. II.
Im ipram ine and prochlorperazine. Comprehensive Psychiatry, 2, 219-226.
W hitm an, R., Pierce, C., Maas, J., & Baldridge, B. (1962). The dreams o f the
experim ental subject. Journal o f Nervous and M ental Disease, 134, 431-439.
W h itty , C., & Lew in, W . (1957). V iv id day-dreaming: A n unusual form o f confusion
follow ing anterior cingulectomy. B rain, 80, 72-76.
W ich lin ski, L. J. (2000). The pharmacology o f threatening dreams. Behavioral and
Brain Sciences, 23, 1016-1017.
W inegar, R. K., & Levin, R. (1997). Sex differences in the object representations
in the dreams o f adolescents. Sex Roles, 36, 503-516.
W inget, C., & Kramer, M . (1979). Dimensions o f dreams. G ainesville: U niversity
o f Florida Press.
W ood, J. S. D., & Dom ino, G. (1989). Do creative people have more bizarre
dreams? A reconsideration. Im agination, C ognition, and Personality, 9, 3 -1 6.
REFERENCES 195
W oolley, J. (1995). The fic tio n a l m ind: Young ch ild re n ’s understanding o f im agina
tio n , pretense, and dreams. Developmental Review, 15, 172-211.
Yamanaka, T ., M o rita , Y., &. M atsum oto, J. (1982). Analysis o f the dream contents
in college students by REM-awakening technique. Folia Psychiatrica et N eurolog'
ica Japonica, 36, 3 3-52 .
Yu, N . (1999). The contemporary theory o f metaphor: The Chinese perspective. P hiladel
phia: John Brightm ans.
Zabriskie, B. (1995). Jung and Pauli: A subtle symmetry. Journal o f A nalytical
Zadra, A . (1996). Recurrent dreams: T h e ir re la tion to life events. In D . B arrett (Ed.),
Traum a and dreams (pp. 231-247). Cambridge, M A : Harvard U n ive rsity Press.
Zadra, A ., & D onderi, D. C . (2000a). Nightm ares and bad dreams: T h e ir prevalence
and relationship to w ell-being. Journal o f Abnorm al Psychology, 109, 273-281.
Zadra, A ., & Donderi, D. C. (2000b). Threat perceptions and avoidance in recurrent
dreams. Behavioral and B rain Sciences, 23, 1017-1018.
Zadra, A ., Nielsen, T ., & D onderi, D. C . (1998). Prevalence o f auditory, olfactory
and gustatory experiences in home dreams. Perceptual and M otor Skills, 87(3,
Pt 1), 819-826.
Zepelin, H . (1972, June). Comparison o f dreams recalled in the laboratory and at home.
Paper presented at the annual m eeting o f the Association fo r the Psychophysio-
logical Study o f Sleep, Chicago.
Zepelin, H . (1980). Age differences in dreams: I. M en’s dreams and them atic
apperceptive fantasy. International Journal o f Aging and Hum an Development,
12, 171-186.
Zepelin, H . (1981). Age differences in dreams: II. D isto rtio n and other variables.
International Journal o f Aging and H um an Development, 13, 37-41.
Zimmerman, W . (1970). Sleep m entation and auditory awakening thresholds. Psy-
chophysiohgy, 6, 540-549.
196 REFERENCES
INDEX
A c tiv a tio n , in p u t, m o d u la tio n ( A IM ) cu ltu ra l v a ria tio n in , 46

m odel gender differences in, 46
characterization o f R E M and N R E M rehearsal theories and, 163-164
dreaming, 149-150 Aggression, m isfortune, negative e m o tio n
A c tiv a tio n -s y n th e s is theory, 147-157, cross-cultural, 26
152 Aggressive actions
bizarreness and, 152-154 ra tin g scales for, 60
c o n tra d ic tio n of, 148-149 Aggressive elem ent
description of, 147 in home dream samples, 45
dream c o n te n t in , 154 A n im a l dream ing, 163-164
dream fo rg e ttin g in, 152 pets, 9 8 -1 0 0
dream ing as psychiatric phenom enon, A ntrobus, J., 19, 152
152 A n x ie ty dreams, 79
dream ing in neonates in , 151 in Freud, 139
h a llu c in a to ry co n te n t in, 148 A rchetypes, 144
na tu ra l scientist’s dream journal A u d ito ry association cortex
connections w ith w aking life vs. in dreaming, 12
w ith b ra in stem, 157 Awareness o f dream ing during dream, 17
consistency in, 155
h profile for, 155-156
study fo r content, 155 Barb Sanders series
study fo r influence o f p o n tin e s tin v baseline sample, 113
u la tio n o n dream features, 155 biographical background, 111
neglect o f co g n itive dim ension in, b lin d spots o f
151, 157 interviewees on, 111-112
neu ro co g n itive model vs., 3 -4 relationships w ith men, 111-112
neuroim aging and lesion studies and, characters and social interactions in
149 aggression/friendly percent in , 115
N R E M dream ing and, 149 percents compared w ith female
o rig in a l, 147 norms, 115
overemphasis o n p o n tin e tegm entum consistency in , 114-115
vs. hypothalam us role, 151-152 d is c o n tin u ity o f dream and w aking
phasic s tim u la tio n fro m bra in stem in, m em ory and past experiences,
147-148 128-131
R E M vs. N R E M in, 148 figurative th o u g h t and, 130-131
revised, a c tiva tio n , in p u t, m odulation guns and rifles, 130
m odel in , 149 horses, 129-130
visual system in, 147 “ stray k itte n s ” and “ lost m en,”
and “ w hole b ra in isomorphism,” 157 128-129
A c tiv ity d is c o n tin u ity o f sexual intim acies w ith
ra tin g of, 59 w a kin g reality, 123-125
Aggression w ish fu l dreams, 123-125
in Barb Sanders, 115, 120, 123 dreamed interactions and w aking co n
hom e and laboratory reports ceptions, 118
differences in , 45 dream le n g th in , 112-113
hom e and laboratory reports o f dreams about fa m ily and friends
197
Barb Sanders series, continued e x tra c tio n o f generic in fo rm a tio n
enactm ent o f conceptions and co n from , 128
cerns about them , 126 inference o f great a m b itio n from ,
dreams o f fa ile d in fa tu a tio n (D erek), 127
1 2 2 -1 2 6 rejections and m isfortunes in , 1 2 7 -
aggressions in , 123 128
forgiveness in , 125 unusual elements
friendly/sexual interactions in , 123, character metamorphoses, 1 31-133
124 com posite characters, 131-132
friends views on, 123 d is tin c tio n s among, 131
in itia tio n of, 122 frequency of, 131
jealousy, 1 2 3 -1 2 4 o bject metamorphoses, 133
resolutions dreams, 125 B arrett, D ., 160, 161
ex-husband dreams (H o w a rd ), 118— Bednar, J. A ., 141
112 Bizarreness
changes in aggression/ffiendliness in a ctiva tio n -syn th e sis theory, 152—
percent over tim e, 120 154
D a rry l and, 118 difference in hom e vs. laboratory re
frequency o f appearances, 119 ports of, 45
in te rv ie w about, 11 8 -1 1 9 ra tin g of, 59
in te rv ie w reflections o n h e r fe e l ra tin g scale fo r, 60
ings, 120-121 B lin d adults
m ixe d feelings about, 119 visual imagery in , 25
negative, 121 B lin d spots
reflective after his death, 121 in Barb Sanders series, 110-112
re p e titio n p rin c ip le in , 122 B rain
tra u m a tic q u a lity of, 122 sites related to dream ing, 11
w is h fu l d im ension of, 122 B ra in stem lesions
H a ll- V a n de Castle system in , 80, dream ing and, 15
8 3 -8 4 B rain s tim u la tio n
h p ro file o f D erek and H ow ard fo r shaping dream co n te n t, 17
dreams, 124 Braun, A ., 1 5 -1 6
intera ctio n s w ith fa vo rite people, B rie f dream diaries
1 1 7 -1 1 8 in c o n ju n c tio n w ith N ig h tc a p m o n ito r
intera ctio n s w ith key people ing at hom e, 50
bro th e r, 117 demand characteristics
fa th e r, 116-117 hasty or confabulated reports and,
G in n y , 117 51
Lucy, 118 p a rtic ip a n t drop o u t rate and, 50
m id d le daughter, 117 tim e to o b ta in dreams, 50
m other, 116
oldest and youngest daughters,
117
patterns of, 116 Castellan, N . J., 64
re lia b ility o f codings for, 113, 114 C a t lesions studies, 15
report d is trib u tio n by years, 112 Characters
theater dreams, 126-128 gender differences in , 62
concerns about perform ance as C h ild re n
themes, 127-128 absence o f dream ing in , 3
continuous w ith w a kin g interests, b lin d , dream imagery in, 23
126 cross-sectional study, 2 1 -2 4
198
age and gender differences in , 21, conceptual system in , 3 0 -3 1
22 C o n tin ge n cy analysis
cognitive structure dream ing, 22 o f H a ll- V a n de Castle coding catego
co n te n t vs. a d u lt co n te n t, 21—22 ries, 7 6 -7 7
p a rticip a tio n in dreams, 22 C o n tin u ity
recall in, 21, 22, 23 in Barb Sanders, 126
visuospatial skills and, 22 between c o n te n t and w aking thought,
developm ental dim ension o f dream ing 26
in, 20 in Emma series, 105
dreams as shared fantasies in, 23, 24 in L u c ille , 110
fro n ta l-lo b e executive fu n ctio n s in vs. Jung’s compensatory fu n c tio n ,
lo n g itu d in a l study, 20-21 145-146
C h o lin e rg ic pathway C o n tin u ity p rin cip le , 2 6 -2 7
in dream generation, 16 c o n n e ctio n o f dreaming and w aking
C h o lin e rg ic system c o g n itio n and, 30
in neurom odulation du rin g R E M C o rtic a l n e tw o rk fo r spatial represen
ha llu cin a to ry imagery and, 150 ta tio n
C o g n itiv e processes in dreaming, 12-13
developm ent o f dream ing and, 4, 5 C u ltu re
C ohen, D., 52 and co n te n t, 26
C o lle c tiv e unconscious, 144
Com pensatory fu n c tio n o f dreams, 145
C onceptual system Disgust
o f dreamer, 32, 33 in anger category o f H a ll- V a n de Cas
experiential categories in , 3 0 -3 1 tle system, 70
in te ra ctio n o f neural structures and Displacem ent
e n vironm ental s tim u li in, in dream -work, 137
3 0 -3 1 D opam inergic system
sensorimotor, 31 in dream generation, 16
spatial relations, 31 in dream ing, 141
expression in dreams, 32 D orsolateral prefrontal cortex
figurative concepts in, 33 in dream ing, 12
C ondensation D ream B ank.net
in dream-work, 137 consistency in dreams and, 101—102
C o n fu sion in c o n te n t analysis
p ro d u ctio n of, 72 com parison w ith n o rm a tive group,
Consciousness 98, 99
developm ent of, 37 d e fin itio n o f categories in , 98, 99
Consistency dreams o f b lin d persons, 101
o f ad u lt content, 3 frequency counts, 98, 99
in content, 27, 28 generation o f percentages, 98, 99
vs. change in re action to co n te n t, 166 scales independent o f H a ll- V a n de
C o n te m p t Castle system, 100-101
in anger category o f H a ll- V a n de Cas sensory references coding scale in ,
tle system, 70 100-101
C o n te n t analysis. See also D ream B ank.net dream series available on, 97
D ream B ank.net for, 9 8 -1 0 2 elements w ith com m on words o r
H a ll- V a n de Castle system of. See also phrases, 9 7 -9 8
H a ll-V a n de Castle system o f Emma series study, 103-105
co n te n t analysis fa c ilita tio n o f H a ll-V a n de Castle c o n
C o n te n t and w aking c o g n itio n te n t analysis, 9 6 -9 8
IN D E X 199
D ream B ank.net, continued free association in , 5 4 -5 5
figurative th o u g h t in dreams and H a ll d e fin itio n of, 57
in c o n ju n c tio n w ith H a ll- V a n de m etaphoric, 5 5 -5 6
Castle coding, 102 o b je ctive and q u a n tita tiv e approach
dreams about bridges, 102-103 of, 57
w edding dreams, 102 ra tin g scales at o rd in a l vs. n o m in a l
large w o rd strings fo r H a ll- V a n de le ve l in, 57
Castle category, 98 ra tin g scales for. See also R a tin g scales
pet anim als study, 9 8 -1 0 0 assumptions in , u n te n a b ility of, 60
findings in , 9 9 -1 0 0 drawbacks to, 5 8 -6 0
male and female norms in , 9 8 -9 9 loss o f in fo rm a tio n w ith , 60
screen v ie w in g w ith , 9 5 -9 6 n o m in a l-e m p iric a l, 58
search engine in , 95 o rd in a l, 58
simple w o rd searches fo r coding social in re la tio n o f c o n te n t to w aking
interactions, 96 conceptions, 60
uses of, 95 re lia b ility /v a lid ity p ro b le m w ith , 59
w ord strings fo r phrases and expres statistical
sions, 96 problems of, 6 3 -6 5
subsamples from , 9 6 -9 7 them atic, 56
Dream c o n te n t th e o re tica l categories in , 57
in a c tiva tio n -syn th e sis theory, 154 D ream c o n te n t studies
in adulthood, 2 5 -2 6 lite ra tu re q u a lity and, 65—66
analysis of, 5 3 -6 3 D ream c o n te n t-w a k in g c o g n itio n rela
changes in and p a rie ta l lobe injuries, tio n sh ip
29 n o n laboratory study
in c h ild re n recallers and, 53
p ersonality dim ensions and, 2 1 -2 2 , D ream ing
24 co n d ition s in, 3 1 -3 2
in college students, 25 co rre la tio n o f changes in w ith w aking
consistency in, 27, 28 c o g n itive defects, 11
consistency th ro u g h o u t n ig h t, 42 and focal lesions, Solms study, 1 0 -15
cross-cultural sim ila ritie s and d iffe r neurophysiological process of, 4 -5
ences in , 26 D ream ing co g n itio n , 1 8 -2 5
drug e ffect on, 2 8 -2 9 Dream journals (dream series)
figurative th o u g h t and, 3 4 -3 5 , 36 Barb Sanders, 80, 8 3 -8 4 , 111-133.
H a ll- V a n de Castle system findings, See also Barb Sanders series
2 5 -2 6 , 27 b lin d spots in, 110-111
in co rp o ra tio n o f e xternal s tim u la tio n com parison w ith laboratory samples
in, 19 and at-hom e systematic awaken
litera tu re on, 65 ings, 49
neurological defects and, 29 D orothea, 49
over R E M periods, 4 1 -4 2 drawbacks w ith , 48
re p e titio n in , 2 7 -2 8 Emma, 1 0 3 -1 0 5 , 110
resemblance metaphors in, 35 freedom fro m investigator influence,
and w a k in g c o g n itio n , 3 0 -3 7 48
w aking consciousness use of, 159—162 h isto ry of, 49
D ream c o n te n t analysis, 53. See also inferences in
Dream c o n te n t interview s w ith people w ho knew
conte n t, 5 7 -6 1 dreamer, 110
developm ent o f categories for, 57 L u c ille series, 1 0 8 -1 0 9
e m p irica l categories in , 5 7 -5 8 L u c ille , 10 9 -1 1 0
200 INDEX
Pauli, 49 D ream S A T, 67, 88, 95, 113
in psychotherapy setting, 5 1 -5 2 D ream series. See Dream journals (dream
studying series)
D ream Bank.net in , 108 D ream theory(ies)
inferences in , 108-111 activation-synthesis, 147-157
use in systematic studies, 49 Freud’s, 136-143
Dream Property Scale fu n ctio n a l, 157-167, 158-165
factors in, 59 fu n c tio n a l psychodynamic, 165-167
Dream reports Jung’s, 144-147
accuracy and representativeness of, D ream -w ork
4 0 -4 6 in Freud, 137
o f c o lle c tio n in laboratory, 44 Dream y states
comparison o f laboratory vs. home in epilepsy, 15, 28
co lle c tio n , 4 3 -4 6 Drugs
in com parison w ith o th e r self' and dream co n te n t, 5
reports, 41 D rug studies
c o rre la tio n w ith actions reported in c o n ju n c tio n w ith neuroim aging
R E M sleep'behavior disorder studies, 2 8 -2 9
actions, 4 0 -4 1 D u ra tio n
co rre la tio n w ith sleep-talking epi recall and, 42
sodes in R EM , 40
freedom fro m cu ltu ra l disto rtio n , 41
analytical methods for, 5 3 -6 1 Elements o f the Past category
o f ch ild re n vs. adults, 2 4 -2 5 o f H a ll- V a n de Castle system, 6 7 -6 8
c o lle c tio n methods for, 4 6 -5 2 Emma series
c o lle c tio n of, 46 appearance o f husband and m in iste r
b rie f diaries, 5 0 -5 1 in
most recent dream method, 4 7 -4 8 c o n tin u ity w ith w aking life , 105
personal dream journals, 4 8 -4 9 frequency of, 103
psychotherapy setting, 5 1 -5 2 separately and together, 104
sleep laboratory awakenings, 47 social interactions w ith , 104-105
co nfabulation in tim e period and frequency, 104
in b rie f dream series, 51 history of, 103
w ith in te n s ity in w aking life inferences in
ra tin g scales for, 5 8 -5 9 d re a m er-theorist disagreement
representativeness o f recallers and, about, 110
5 2 -5 3 E m o tio n a l dreams
in sleep laboratories fro m n ig h t awak dream solutions vs. w aking realizations
enings, 40 and, 159
statistical analysis of, 6 3 -6 5 problem -solving fu n c tio n of, 159
u n it o f analysis in, 6 3 -6 5 E m otional knowledge
DreamResearch.net, 67, 85, 87, 113 in te n sity dim ension of, 72
D re a m S A T spreadsheet on, 67, 88, po sitive -n e g a tive dim ension of, 72
95, 113 Em otions
SearchCodings program on, 76, 98, in Freud’s dream theory, 139
126 E m p irica l ra tin g scales
Dreams dimensions of, 58
conceptions and concerns in , 33 Epilepsy
personal nature of, 32 dreamy states in , 15, 28
as sim ulations o f w a kin g w orld, 168 nightmares in, 28
types of, 79 E xiste n tia l dreams, 79
IN D E X 201
Figurative th o u g h t F u n ctio n a list psychodynamic theories,
in Freud, 143 165-167
lin k w ith dream co n te n t, 3 3 -3 4 , 35, K ram er theory o f dream c o n ta in m e n t
66 o f e m o tio n a l surge, 166-167
Fisher, S., 143b w o rk in g through traum atic experi
F lying dreams, 148 ences, 165
as prim ary m etaphor, 3 3 -3 4 F u n c tio n a l theory(ies), 157-167
Food and E ating category adaptionist assumption and, 158
o f H a ll- V a n de Castle system, 6 7 -6 8 in fo rm a tio n o n physical illness, 159
Foulkes, D., 18, 19, 21, 23 in sp ira tio n a l, 162
Free association problem -solving theory, 158-162
con stru ctio n o f “ m eaning” problem rehearsal, 163
w ith , 54 o f unrecalled dreams, 162—165
description of, 54
in in te rp re ta tio n w ith in psychothera-
peutic process, 54 G ender
problems o f re lia b ility and generaliz- o f characters, cross-cultural, 26
a b ility in , 5 4 -5 5 G ender differences
shaping th ro u g h suggestion, persua fa ilu re to c o n tro l o r correct, 62
sion, conversion, 54 G oodenough, C ., 53
usefulness of, 55 G ood fo rtu n e /M isfo rtu n e categories
Freud, S., 136, 137, 139, 140, 141 death vs. re tu rn fro m death, 78
Freud’s theory o f dreams e n v iro n m e n ta l obstacles/“ b o u n tifu l e n
anxie ty dreams and, 139 v iro n m e n ts,” 7 8 -7 9
ch ild re n in, 136 fa llin g o r danger o f falling/suspension
co g n itive processes in , 137 o r flying, 78
c ritic is m fro m researchers tra in e d sub lost, damaged, defective/m iraculous
lim in a l s tim u la tio n , 138 events, magical objects, 78
dream -w ork in , 137 suffering/recovery, 78
em otions and, 139 threats to w ell-being/enhancem ent o f
as guardians o f sleep w ell-being, 78
re visio n of, 166 Greenberg, R., 143
as “ guardians o f sleep,” 142 Gregor, Thom as, 32
in fa n tile sexual desires in , 1 4 1 -1 4 2
la te n t dream -thoughts and, 143
loss o f dream ing in , 142 H a ll, C ., 35, 57, 70, 72, 77
memories in shaping c o n te n t, 136 study o f c o n te n t inside and outside
and m ethod o f free association, 143 laboratory, 4 3 -4 6
overcom ing resistance vs. persuasion aggression in , 45
and conversion, 143 bizarre elem ent in, 45
as p ro d u ct o f fig u ra tive th o u g h t, differences in, 4 4 -4 5
143 dram atic in te n sity index and, 45
regard fo r representability, 137 s im ila rity in, 44
re p e titiv e dreams o f posttraum atic H a llu c in a to ry imagery
stress disorder and, 140-141 ch o lin e rg ic n e urom odulation d u rin g
repression in dream fo rg e ttin g , 141 R E M and, 150
secondary re visio n in, 137 H a ll- V a n de Castle coding system,
sublim in a l s tim u la tio n studies in de 6 7 -9 4
fense of, 137—138 categories o f
traum a and, 139, 140 added, 6 7 -6 8
w ish fu lfillm e n t in, 136, 137, 143 o rig in a l, 67
202
c o m b in a tio n o f n o m in a l e m pirical cat random ization strategies in
egories of, 74 approxim ate random ization, 8 5 -8 6
c o m b in a tio n o f theoretical ra tin g bootstrapping, 8 7 -8 8
scales and, 75 u tilit y of, 85
co m p u ta tio n o f re lia b ility in significance testing
percentage-of-agreement in, 6 8 -6 9 o f social in te ra c tio n indexes, 86
c o n te n t indicators in statistical analysis of, 8 4 -8 6
ca lcu la tio n formulas for, 70, 71 bootstrapping fo r confidence in te r
c o rre c tio n fo r report le n g th variations, vals, 8 7 -8 8
7 9 -8 4 effect size in , 8 8 -9 0
d escription of, 4, 5, 61 nonparam etric statistics and, 84
em otions in proportions o r c h i square and,
categories of, 70 8 4 -8 5
indicators in p values and, 84, 8 5 -8 6
ca lcu la tio n of, 70, 71 H a rtm a n n , E.
inte rco d e r re lia b ility of, 68 dream as w o rkin g through traum atic
mem orable dreams and, 7 7 -7 9 experiences, 165-166
n o m in a l categories Hobson, J., 15, 147, 148, 150, 151, 152
expression in c o n te n t indicators, h S tatistic
percentages and ratios, 69 c a lc u la tio n of, 88
n o m in a l measurement vs. ra tin g scales in d e te rm in a tio n o f effect sizes, 8 8 -9 0
in in graphic representation, 8 9 -9 0
n o rm a tive findings o n college m en fo r independent and repeated mea
and women, 72—76 sures, 89
comparisons by gender, 73 and mean effect size, 89
group vs. in d iv id u a l comparisons in profile fo r H a ll- V a n de Castle n o r
of, 74 m ative sample, 9 0 -91
replications of, 72, 74 use w ith percentage data, 88
reports needed fo r significant p values,
9 2 -9 3
sample sizes Inappropriate dress in p ublic
h difference to fin d statistically sig w aking m e ta phor-dream linkage,
n ific a n t difference, 93, 94 34
sample sizes for, 9 2 -9 4 In sp ira tio n a l dreams, 162
sequences and connections in , 7 6 -7 7 In te n sity
social interactions recall and, 42
n o m in a l categories for, 69
rates per character analysis of, 69
H a ll- V a n de Castle data Joy
approxim ate random ization and, 85 correspondence w ith happiness in
frequency d istrib u tio n , H a ll- V a n H a ll- V a n de Castle system, 70
de Castle categories, 8 5 -8 6 Jung, C .
confidence intervals in analysis o f P auli series, 49
bootstrapping strategy for, 8 7 -8 8 his dream as structural diagram o f h u
h statistic m an psyche, 145
in d e te rm in atio n o f effect sizes, Jung’s theory o f dreams, 144-147
8 8 -9 0 a m p lifica tio n m ethod and sym bolic in
p values fo r percentage differences be te rp re ta tio n in , 55
tw een samples archetypes in , 144
approxim ate random ization and, changes in c o n te n t fo r m id life in d iv id
8 5 -8 6 u a tio n and integration, 146
IN D E X 203
Jung’s th e o ry o f dreams, continued Masochism scale
consistency vs., studies o f, 1 4 6 - H a ll- V a n de C astle system and,
147 7 4 -7 5
c o lle c tiv e unconscious in , 144 M a tu rity assessment
c ritiq u e of, 144 H a ll—V a n de Castle system in , 75, 76
com pensatory fu n c tio n in, 145 M cC arley, R., 147, 148
c o n tin u ity p rin c ip le vs., 145-146 M em orable dreams, 77—79
sym bolic in te rp re ta tio n s as m etaphoric coding patterns for, 79
analyses in , 144-145 H a ll- V a n de C astle system and
good fo rtu n e categories in , 7 8 -7 9
“ good fo rtu n e ” in , 77
M em ory schemata
Kram er, M ., 6 5 -6 6 , 165-166, 167
in dreaming, 32
M e ta p h o ric analysis
co lla b o ra tio n o f dreamer and in te r
Lengths preter in , 55
Barb Sanders series study o f problems o f re lia b ility and generaliz-
H a ll—V a n de Castle system in , 80, a b ility in , 56
8 3 -8 4 sym bolic in te rp re ta tio n and, 5 5 -5 6
co rre c tio n o f variations in , 7 9 -8 4 M etaphor(s) in dreams, 3 3 -3 4
c o n te n t indicators in, 80 o f b lin d persons, 101
percentages and ratios in , 7 9 -8 0 conceptual, 3 4 -3 5
co rre la tio n o f category frequencies fo r personal fro m past experience, 36
w ord count, 8 0 -8 1 resemblance, 35
c o n te n t in d ic a to r c o n tro l of, 82 M etonym y, 35
fa ilu re to c o n tro l for, 62 M ood
m in im u m and m a xim u m fo r analysis, recall frequency and, 52
6 2 -6 3 M ost recent dream m ethod
strategies to c o n tro l fo r v a ria tio n in, in group setting o n standardized form ,
62 47
Leucotom y problems
and loss o f dream ing, 13 w ith c h ild re n , 47
Loss o f dream ing lack o f personality o r c o g n itive
bra in sites in , 11, 13 measures w ith , 47
parietal lobe in , 24 standardized H a ll- V a n de C astle form
Loss o f R E M for, 48
b ra in sites in , 11, 15 vs. dream diary, 51
Love
correspondence w ith happiness in
H a ll- V a n de Castle system, 70
Lucid dream ing N e u ra l n e tw o rk
fro n ta l co rte x a c tiv ity during, 18 agreement o n contours of,
R E M sleep and, 1 7 -1 8 10
L u c ille series co g n itive processing areas in , 11 6 -1 7
c o n tin u ity w ith w a kin g life , 110 disagreement o f theorists, 10
exception to, 110 neuroim aging studies, 9, 10
inferences in N e u ro co g n itive m odel
d re a m e r-th e o ris t disagreement absence and excess o f dream ing and,
about, 110 168
social in teractions w ith m a in charac coherence and meaningfulness o f
ters, 10 9 -1 1 0 dreams and, 169
IN D E X
conceptual systems in forebrain and, P a rtic ip a tio n
169 in c h ild re n , 22, 37
c o n te n t in , 2 5 -3 0 index o f sense o f self, 37
and w aking co g n itio n , 3 0 -3 7 Pauli, W olfgang
dream fu n c tio n and nondream ing in dream journals o f and Jung’s analysis,
ch ild re n and adults w ith b ra in le 49
sions, 168 Percentage indicators
dream ing as c o g n itive process and, in comparison o f R E M and N R E M re
1 69-170 ports, 63
dream ing c o g n itio n in , 18-25 Positron emission tom ography (P E T), 9
in te g ra tio n w ith developm ental stud o f lu c id dreaming, 18
ies and co n te n t analysis, 167 P roblem -solving theory
lu c id dream ing and, 1 7 -1 8 in C a rtw rig h t study o f divorce, 159,
neural n e tw o rk in , 9 -1 7 160
sim ilarities between dream ing and wak recaller and nonrecallers and, 1 5 8 -
ing co g n itio n in , 168 159
te sta b ility of, 169 re fu ta tio n of, 1 58-159
N euroim aging studies s o lu tio n in w aking re fle ctio n o n
R E M and N R E M stages in, dream, 161, 162
10 in study o f college students, 160-161
N europsychological in fo rm a tio n Psychotherapy
Solms study o f dream ing and focal dream reports in
b rain lesions, 1 0 -1 5 demand characteristics o f therapeu
N ig h tca p tic re la tio n sh ip in, 51
sleep m o n ito r, 49 dream journals in , 5 1 -5 2
N ightm ares focus on problems vs. dreams, 51
b ra in sites in, 11, 14 m ost recent dream m ethod in, 51
w ith epilepsy, 15, 28 m etaphoric interpretations o f dreams
new endings for, 37 in, 37
in posttraum atic stress disorder, 27, 28
N o m in a l categories
conversion to percentages and ratios, R andom ization strategies
63 approxim ate, 8 7 -8 8
in H a ll- V a n de Castle coding system, bootstrapping, 8 7 -8 8
61, 63 u tility of, 85
N o m in a l scales R a tin g scales
re lia b ility of, 61 applications of, 5 8 -5 9
vs. ra tin g scales, 61 fo r co n te n t analysis, 5 8 -6 9
fo r dimensions o f dream salience, 59
fo r e m o tio n a lity dim ension, 59
O rd in a l scales, 58 in re la tio n o f dream ing to neurophysi
O rd in a ry dreams, 79 ology o f sleep, 60
R ecall
in ch ild re n , 21, 22, 23
Parietal lobe dream d u ra tio n and, 42
c o n te n t and, 30 recency and d u ra tio n in , 42
in dream ing, 24 Recallers
visuospatial skills and, 24 c o g n itive variables in, 52
Parietal regions com parison w ith nonrecallers, 52
in sym bolic and spatial mechanisms, interest in dreams and, 5 2 -5 3
13 personality variables in, 52
IN D E X 205
Recallers, continued S im ulations
physiological factors in , 52 in dream ing, 32
representativeness of, 5 2 -5 3 Sleep laboratory awakenings
R ecurrent dreams, 28 problems fo r participants and research
Regard fo r representability ers, 47
in dream -work, 137 representativeness advantage to, 47
Rehearsal theories Sleep onset dreams, 150
re la tio n to th re a t and aggression, Solms, M ., 13, 15, 16
163-164 Solms study
R E M and dream ing same in a ll m am o f d re a m in g -n e u ro lo g ica l structure re
mals, 163 la tio n ship , 1 0 -15
REM findings
dream ing in absence o f, 15 b ra in areas in a ctive d u rin g R E M
R e p e titio n and N R E M , 12
consistency of, 27 b ra in areas in vividness and w ak
drug and neuroim aging studies and, ing in tru sio n o f dreams, 14
2 8 -2 9 b ra in stem lesions and dream ing, 15
em o tio n a l b ra in in , 5 injuries in visual association cortex
th e m a tic p o in t in , 3 4 -3 5 and changes in dream ing, 12
R e p e titio n p rin c ip le parallel o f visual im pairm ents in
and nightm are in posttraum atic stress dream ing w ith w a kin g deficits,
disorder, 27, 28 12
recu rre n t dreams and, 27, 28 tem poral lobe in ju ry and n ig h t
frequency of, 27 mares, 14-15
repeated themes and, 28 v e n tro m e d ia l lesions in loss o f
R e p e titiv e nightm ares dream ing, 13
in posttraum atic stress disorder S patial mechanisms
om ission o f Freud, 140-141 b ra in site for, 13
as re fu ta tio n o f Freud’s wish- S p o tlig h t dreams
fu lfillm e n t theory, 140-141 in th e m a tic analysis, 56
R e p lica tio n studies S ta tis tic a l analysis
im portance of, 90, 92 c o m b in a tio n o f nonparam etric statis
Resemblance metaphors, 35 tics and random ization strategies
in dreams, 35 in , 54
R o th , T ., 6 5 -6 6 o f H a ll- V a n de Castle data
replications studies and, 90, 92
nonparam etric statistics in , 64
param etric tests and, 63—64
Salience problem s of, 6 3 -6 5
bias in hom e dream reports, 45 p values and, 64
dim ensions of, 59 sample size fo r statistical significance,
Salis, P., 6 5 -6 6 65
Secondary revision sm all sample size issue in , 6 4 -6 5
in dream -w ork, 137 Steriade, M ., 151
S ensorim otor cortex Stress
in dream ing, 12 recall frequency and, 52
S equential analysis S u b lim in a l stim u la tio n
H a ll- V a n de Castle coding categories in defense o f Freud’s th e o ry o f dreams,
and, 76 137
S hevrin , H ., 138 S u b lim in a l s tim u li
Siegel, S., 64 and Freud’s theory o f dreams, 138
IN D E X
Surprise V a n de Castle, R ., 70, 72, 77
correspondence w ith confusion in V en tro m e d ia l lesions
H a ll—V a n de Castle system, 70, loss o f dream ing and, 11, 13
72 V en tro m e d ia l region
S ym bolic mechanisms in loss o f dream ing, 13
b ra in site for, 13 V ig ila n c e -fe a r system
as neurocognitive record o f traumas,
tensions, 28
T e m p o ra l-lim b ic region V isual association cortex
in dreamy states, 28 in dreaming, 12
T em poral lobe in ju ry V isual imagery
nightm ares and, 14-15 defects in
T h e m a tic analysis bra in sites in, 11, 12
description of, 56 Visual im a g in a tio n
problems o f re lia b ility and generaliz- in children, 2
a b ility in , 56 Visuospatial a b ility
T h e o re tic a l ra tin g scales in recallers, 52
co n stru ctio n of, 58 Visuospatial skills
Thom pson, N ., 158 in ch ild re n , 22, 23
Transcendent dreams, 79 Vividness
T raum a ra tin g of, 59
in re la tio n to dreams. See R e p e titive
dreams
in Freud, 140
W a k in g co g n itio n
parallels w ith dreaming, 6, 20
Unconscious th o u g h t W a kin g dreaming
transform ation in to dreams, 137 b ra in sites in , 11, 14
U n its o f analysis W a k in g th o u g h t-d re a m in g parallel,
n o m in a l categories in 1 9 -20
conversion o f frequencies in to per W is h -fiilfillm e n t
centages and ratios, 63 in Freud’s theory o f dreams, 136-137
U n re ca lle d dreams defense of, 141
fu n c tio n of, 162-165 W o o lle y, J „ 22, 24
clearing useless recent memories, 163
memory consolidation, 163
IN D E X 207
ABOUT THE AUTHOR
G . W illia m D o m h o ff received h is B A a t D uke U n iv e rs ity , h is M A a t K e n t

S tate U n iv e rs ity , and h is P h D a t th e U n iv e rs ity o f M ia m i— a ll in psychology.
H e has ta u g h t a t th e U n iv e rs ity o f C a lifo rn ia , S anta C ruz, since 1965, w here
he is n o w a research professor. H e is th e a u th o r o f The M ystiq u e o f D ream s
(1 9 8 5 ) and F in d in g M e a n in g in D ream s (1 9 9 6 ) as w e ll as num erous jo u rn a l
a rticle s o n dream c o n te n t.

The Scientific Study of Dreams 2003

Uploaded by

Copyright:

Available Formats

The Scientific Study of Dreams 2003

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Scientific Study of Dreams 2003

Uploaded by

Copyright:

Available Formats

TheScientific Study of

F irst P rin tin g , N ovem ber 2002

T o order In the U .K ., Europe, A fric a , and the M iddle

Typeset in Goudy by W o rld C om position Services, In c., S terlin g, V A

P rinter: P ort C ity Press, Baltim ore, M D

L ib ra ry o f Congress C atalog ing -in -P ub lica tion D ata

B ritis h L ib ra ry C atalog uin g-in-P u blicatio n D ata

Printed in the U nited States o f A m erica

Acknowledgments ................. ................ .....................................

Introduction: The N eurocognitive Approach to Dreams .... ............... 3

Chapter 1. Toward a N eurocognitive M odel o f Dreams .. ............... 9

Chapter 2. M ethodological Issues in the Study

Chapter 3. The H a ll-V a n de Castle System .... ............... ............. 67

Chapter 4. A New Resource fo r C ontent Analysis .......... ............. 95

Chapter 5. New Ways to Study M eaning in Dreams ....... ........... 107

Chapter 6. A C ritique o f T raditional Dream Theories .... ........... 135

References .... ......................................... ............. .................. ...... ........... 171

Index ........... ........................................... . .. ......... ............ ........... 197

Thanks to David Foulkes fo r two careful readings o f the entire manu­

B uilding on established descriptive findings in three separate areas o f

■ Dreaming occurs when a relatively specific neural netw ork—

4 THE SCIENTIFIC STUDY OF DREAMS

These generalizations lead to several hypotheses concerning the links

6 THE SCIENTIFIC STUDY OF DREAMS

INTRODUCTION: THE NEUROCOGNITIVE APPROACH TO DREAMS 7

8 THE SCIENTIFIC STUDY OF DREAMS

The neurocognitive model o f dreaming and dreams proposed in this

Research in neuropsychology and neuroimaging converged in the late

TOWARD A NEUROCOGN1TNE MODEL OF DREAMS 11

TOWARD A NEUROCOGN1T1VE MODEL OF DREAMS 13

TOWARD A NEUROCOGN1TNE MODEL OF DREAMS 15

The phenomenon o f becoming aware o f a dream w hile it is ongoing

TOWARD A NEUROCOGNITIVE MODEL OF DREAMS 17

The serendipitous discovery o f REM sleep in 1953, especially the

THE SCIENTIFIC STUDY OF DREAMS

Probably the most general conclusion to be reached from a wide variety

However, on the rare occasions when stim u li are incorporated, “ the

TOWARD A NEUROCOGNITIVE MODEL OF DREAMS 19

THE SCIENTIFIC STUDY OF DREAMS

TOWARD A NEUROCOGNITIVE MODEL OF DREAMS 21

THE SCIENTIFIC STUDY OF DREAMS

24 THE SCIENTIFIC STUDY OF DREAMS

A lth o ug h there are several systems o f content analysis th a t have made

TOWARD A NEUROCOGNITIVE MODEL OF DREAMS 25

THE SCIENTIFIC STUDY OF DREAMS

TOWARD A NEUROCOGNITIVE MODEL OF DREAMS 27

28 THE SCIENTIFIC STUDY OF DREAMS

TOWARD A NEUROCOGNITIVE M ODEL OF DREAMS 29

Findings fro m the study o f dream co n te n t suggest lin ks n o t o n ly w ith

30 THE SCIENTIFIC STUDY OF DREAMS

32 THE SCIENTIFIC STUDY OF DREAMS

TOWARD A NEUROCOGNITIVE MODEL OF DREAMS 33

THE SCIENTIFIC STUDY OF DREAMS

TOWARD A NEUROCOGNIT1VE M ODEL OF DREAMS 35

36 THE SCIENTIFIC STUDY OF DREAMS

TOWARD A NEUROCOGNITIVE MODEL OF DREAMS 37

38 THE SCIENTIFIC STUDY OF DREAMS

T his chapter examines the m ajor m ethodological issues th a t arise in

T H E A C C U R A C Y A N D R EPR ESENTATIVEN ESS

O ne o f the m ost curious and least understood facts about dreams is

Thanks to David Foulkes fo r two careful readings o f the entire manu

T he m ost objective and structured co n te xt fo r the e fficie n t and in e x

D issatisfaction w ith the re lia b ility and generalizability o f free associa

1. E m otion ality/evaluatio n, as indexed by pairings such as re

It should be obvious that a mean and standard deviation may be com

T h e coding categories fo r em otions may prove to be p a rticu la rly im p o r