ArticlePDF Available

Imaginary Companions, Creativity, and Self-Image in Middle Childhood


Abstract and Figures

This study investigates 4 questions: First, whether there is a relationship between imaginary companions and creative potential; second, whether children with negative self-images are more likely to have imaginary companions; third, whether there are gender differences among those children who have imaginary companions; and, finally, what aspects of imaginary companions and what characteristics of those who invent them are related to creativity. The measurements used were a questionnaire about imaginary companions, 3 estimates of creative potential, and a self-image inventory. Among the 69 participating 4th graders, 52% reported having (had) imaginary companions. The children with imaginary companions were more creative on 2 of 3 estimates of creativity and had lower self-image scores. The self-image differences were greatest on the subscales measuring psychological well-being and peer relations. It was more common for girls to have imaginary companions. Aspects associated with creativity among the children with imaginary companions were, for example, elaboration of the companion's character and number of imaginary companions.
Content may be subject to copyright.
Imaginary Companions, Creativity, and Self-Image
in Middle Childhood
Eva V. Hoff
Lund University
ABSTRACT: This study investigates four questions:
First, whether there is a relationship between imagi-
nary companions and creative potential; second,
whether children with negative self-images are more
likely to have imaginary companions; third, whether
there are gender differences among those children who
have imaginary companions; and, finally, what aspects
of imaginary companions and what characteristics of
those who invent them are related to creativity. The
measurements used were a questionnaire about imagi-
nary companions, 3 estimates of creative potential,
and a self-image inventory. Among the 69 participating
4th graders, 52% reported having (had) imaginary
companions. The children with imaginary companions
were more creative on 2 of 3 estimates of creativity and
had lower self-image scores. The self-image differ-
ences were greatest on the subscales measuring psy-
chological well-being and peer relations. It was more
common for girls to have imaginary companions. As-
pects associated with creativity among the children
with imaginary companions were, for example, elabo-
ration of the companion’s character and number of
imaginary companions.
Research on imaginary companions1began just prior
to the turn of the last century (Vostrovsky, 1895). At
that time, imaginary companions were believed to be
signs of mental illness. In the 1940s, there was renewed
interest in the topic, this time within the
psychodynamic tradition (Ames & Learned, 1946;
Bender & Vogel, 1941); this interest continued for
many decades. The phenomenon of imaginary com-
panions has been discussed in terms of psychoanalytic
concepts such as the superego, the ego, and the id
(Bach, 1971; Nagera, 1969; Sperling, 1954). Imagi-
nary companions were also perceived as a manifesta-
tion of different kinds of defense mechanisms, for ex-
ample the companion could be described as an effect of
projection, when a child’s unfavorable characteristics
were projected onto the pretend playmate (Bach, 1971;
Nagera , 1969; Sperling, 1954). In these psychoana-
lytic studies, children with imaginary companions
were reported as having personality defects. However,
M. Taylor (1999) pointed out certain problems with
many of these early studies. Among other things, the
children were not randomly selected, but recruited
from clinics and hospitals where individuals were par-
ticularly likely to have psychosocial and emotional
The reported frequency of children with imaginary
companions varies across different studies. Early stud-
ies showed that 13%–31% of children had make-be-
lieve friends (Ames & Learned, 1946; Hurlock &
Burstein, 1932; Svendsen, 1934), but later research has
uncovered greater figures. Singer and Singer (1992)
found an incidence of 65% in a sample of 111 children,
and M. Taylor (1999) 63% among 100. One reason for
Creativity Research Journal
2005, Vol. 17, No. 2 & 3, 167–180
Copyright © 2005 by
Lawrence Erlbaum Associates, Inc.
Creativity Research Journal 167
Eva Hoff was on a postdoctoral fellowship during 2004 at Brandeis
University, Boston, MA.
I would like to express my gratitude to the children who partici-
pated in this study and to their teachers who provided time for me to
meet the children. I would also like to thank my advisor Ingegerd
Carlsson for her support and Gun Persson for her co-assessment. The
staff of the division of developmental psychology at Lund University
is thanked for their general advice regarding the article.
Correspondence and requests for reprints should be sent to Eva
Hoff, Department of Psychology, Lund University, Box 213, SE-221
00 Lund, Sweden. E-mail:
1In the text, the terms “imaginary companions,” “make-believe
friends,” and “pretend playmates” will be used interchangeably to
denote this phenomenon.
differences in reported frequencies of the phenomenon
is that in some studies (Ames & Learned, 1946; Bender
& Vogel, 1941; Bouldin & Pratt, 1999; Manosevitz,
Prentice, & Wilson, 1973) the children’s parents were
the informants. In other studies the children’s own ac-
counts were used. M. Taylor (1999) interviewed both
children and their parents and demonstrated that the
parents of younger children were in most cases aware
of their children having imaginary companions,
whereas parents of older children were not. Only 20%
of parents of 6- and 7-year-olds knew about their chil-
dren’s pretend playmates. Another reason for different
incidences in earlier research might be that the defini-
tion of the phenomenon varies. Some researchers in-
clude anthropomorphized dolls and others do not.
Svendsen (1934) only included invisible make-believe
friends in her definition. Singer and Singer (1992) ar-
gued that dolls, teddy bears, and other objects can be
included if they assume humanlike properties. But they
did not regard dolls and stuffed animals that were sim-
ply carried around, or treated concretely as transitional
objects, as imaginary companions. Singer and Singer’s
definition was adopted in this study.
Different scholars have concentrated their studies
on different age groups. There are those who maintain
that imaginary companions are most common in the
preschool period with a peak at 4 years (see M. Taylor,
1999, for an overview), and others who contend that
the phenomenon is equally frequent during the early
school years (Hurlock & Burstein, 1932; M. Taylor,
1999). Singer and Singer (1992) proposed that, even if
imaginary companions disappear in their most primi-
tive form between 6 and 8 years, “the process of peo-
pling one’s private thoughts with companionable
souls” (p. 110) continues throughout life.
Among the existing research, very little attention
has been directed toward middle childhood. Most stud-
ies investigating imaginary companions have focused
on preschool children’s present playmates, or on retro-
spective accounts of adolescents and adults. Many
children retain their make-believe friends during mid-
dle childhood and beyond, some until the age of 18
(Seiffge-Krenke, 1997; M. Taylor, 1999). This study
focuses on the middle childhood age group.
Different correlates of having imaginary compan-
ions are discussed in the research literature.
Seiffge-Krenke (2000) highlighted four possible rela-
tionships apart from the association with
psychopathology discussed above: the giftedness hy-
pothesis, the deficit hypothesis, the narcissism
hypothesis, and the impulse control hypothesis. The
first one will be dealt with in some detail in the next
section, and the last three briefly in the following. The
deficit hypothesis holds that pretend playmates can
compensate for minor deficiencies in the child’s envi-
ronment (Harter & Chao, 1992), such as feelings of
loneliness or of being neglected. The narcissism hy-
pothesis holds that children with make-believe friends
are especially unwilling to give up the egocentrism of
early childhood (something also discussed in Hoff,
2005), perhaps due to their having experienced narcis-
sistic blows, such as abuse or the death of family mem-
bers. Finally, the impulse control hypothesis supposes
that children use their pretend playmates for ego sup-
port (or as a superego) during a transitional phase on
their way to autonomous self-regulation.
Imaginary Companions and Creativity
Some scholars have proposed that children with
imaginary companions are more gifted, or more pre-
cisely, that imaginary companions are precursors of
creativity (Myers, 1979; Singer & Singer, 1992).
Myers presented six case studies showing that individ-
uals who had had imaginary companions as children
exhibited creative capacity as adults, and in some
cases, the imaginary companions inspired adult cre-
ativity in different ways. Somers and Yawkey (1984)
provided detailed examples of how children’s intellec-
tual and creative growth can benefit from having pre-
tend playmates. They maintained that make-believe
friends have a connection to creative growth regarding
children’s elaboration of such friends’ characteristics
and roles. Through interaction with imaginary com-
panions, children practice and expand creative thought.
They discover opportunities, explore materials, and
use them in new ways in their play with their pretend
playmates. According to Somers and Yawkey, imagi-
nary companions promote originality of ideas.
Mackeith (1982) elucidated another kind of imagi-
nary play in which imaginary companions are in-
cluded, namely imaginary worlds, so-called
paracosms. This pretend play comprises elaborated
imaginary worlds that the children have developed
and for which the children have sometimes invented
168 Creativity Research Journal
E. V. Hoff
special peoples, local societies, countries with special
architecture, species of flowers and trees, and lan-
guages (Cohen & Mackeith, 1991; Mackeith, 1982;
Singer & Singer, 1992). An interesting question is
whether children who also invent paracosms are more
creative than those who only have “ordinary” pretend
Smith and Carlsson (1990) hypothesized that
make-believe friends constitute a sign of creative po-
tential. However, systematic investigations are rare.
One study of 800 high school pupils, which investi-
gated the association between adolescents’ creativity
and retrospective reports of their having had imaginary
companions as children, indicated such a relationship,
particularly for literary creativity (Schaefer, 1969). In
Manosevitz, Fling and Prentice’s (1977) attempt to
replicate Schaefer’s results, however, no creativity dif-
ferences were found between 5-year-olds (84 children
in total) who had pretend playmates and those who did
not. In another study on 40 children, highly imagina-
tive children scored higher on creativity than did their
less imaginative peers (Singer, 1961). To decide
whether the children were highly involved in fantasy,
one of the questions asked was whether they had an
imaginary companion.
Singer (1961) argued that the ability to fantasize can
be seen as a cognitive style or even a skill. In his study,
41 children (6–9 years) were divided into two groups:
one high imaginative and one low imaginative. The
groups differed on many variables, for example, style
of defense, identification patterns, waiting ability, and
creativity. Among other things, the highly imaginative
individuals had experienced periods of relative soli-
tude that had given them opportunities for practicing
fantasy play and daydreaming, something that, accord-
ing to Singer, constitutes a dimension of human com-
Creativity could be defined as a productive or gen-
erative novel way of experiencing reality—including
the perceiver’s own self (Hoff & Carlsson, 2002, p. 22).
This definition emphasizes the individual’s way of ex-
periencing and way of being. Other definitions stress
the product aspect of creativity—that it should result in
new, original, useful products that actually come to use
(Martindale, 1989). But what is a product? Are imagi-
nary companions useful and novel products? For those
children who invent them, they probably are, but per-
haps not according to the constructors of those defini-
tions. In this study, imaginary companions were con-
ceptualized as a way of experiencing the world in a
productive or generative novel way.
Imaginary Companions and Social
Some scholars assert that children with imaginary
companions are not as socially competent as children
without such companions (Bouldin & Pratt, 1999;
Harter & Chao, 1992). However, other scholars have
claimed that the presence of imaginary companions
may indicate that children are socially well adjusted
(Manosevitz et al., 1973; Singer & Singer, 1992; M.
Taylor, 1999). In the study by Singer and Singer, chil-
dren with imaginary companions showed more posi-
tive emotionality and were less overtly aggressive dur-
ing play compared to children without imaginary
companions. In another study, children with make-be-
lieve friends were described by their parents as happier
in day-to-day activities and as more verbally commu-
nicative (Manosevitz et al., 1973). These children gave
up playing with the imaginary companion when real
playmates appeared. There are also studies indicating
that make-believe friends facilitate children’s cognitive
development. For example, theory of mind was better
developed in young children with pretend playmates
(M. Taylor & Carlson, 1997). The explanation given
was that these children had practiced understanding
other people’s minds through their play with imaginary
Children without siblings as well as first-borns have
imaginary companions to a greater extent, according to
some scholars (Bouldin & Pratt, 1999; Hurlock &
Burstein, 1932; Gleason, Sebanc, & Hartup, 2000;
Manosevitz et al., 1973; M. Taylor, 1999). However, in
several of these studies, parental reports were used to
obtain data about the pretend playmates. Thus, an al-
ternative explanation could be that parents mainly no-
ticed imaginary companions in only or first-born chil-
dren, and/or that younger siblings’ imaginary
companions were kept secret to a larger extent. Particu-
larly in earlier studies, only children with make-believe
friends were described as shy or even socially deficient
(Ames & Learned, 1946).
Very few studies have been conducted where chil-
dren’s self-images have been used to investigate differ-
Creativity Research Journal 169
Imaginary Companions, Self, and Creativity
ences between children with and without imaginary
companions. In this study, the question of how children
with imaginary companions look upon themselves was
Imaginary Companions and Gender
There are gender differences in earlier research.
Many scholars have demonstrated that girls are more
likely to have imaginary companions (see M. Taylor,
1999, for an overview). How can this be explained? If
fantasy play in general is considered, there is no evi-
dence for gender difference. One suggestion is that
girls and boys play differently with imaginary charac-
ters (M. Taylor, 1999). Boys more often impersonate
cartoon or film characters than do girls. Boys become
“Superman,” they do not socialize with him.
One gender difference revealed in Singer and
Singer’s (1992) study is that boys more frequently have
animal friends than do girls, whereas girls have human
pretend playmates to a greater extent. Additionally,
girls are more likely to have imaginary companions of
the opposite sex. In their study, 42% of the girls had at
least one pretend playmate of the opposite sex,
whereas only 13% of the boys had a female make-be-
lieve friend. In another study, girls’ imaginary compan-
ions were found to be less competent than their inven-
tors, whereas boys had companions that were admired
and idealized (Harter & Chao, 1992). Girls’ compan-
ions were protégés whereas those of boys were heroes
or idols. According to the authors, this might reflect
two different mechanisms for handling issues of mas-
tery and competence.
Aims of the Study
The aims of this study were to scrutinize the rela-
tionship between the presence of imaginary compan-
ions and variables such as creativity, self-image, and
gender. One question was whether imaginary compan-
ions might be an expression of a creative disposition. A
second question was whether children with make-be-
lieve friends had negative self-images. A third question
involved discovering which aspects of the imaginary
companionship were especially creative. A final ques-
tion addressed possible gender differences in the phe-
The participants were 69 fourth graders (35 girls
and 34 boys) in six classes at three Swedish schools,
with somewhat different demographic profiles, but
with the majority from middle class homes. In two of
the classes only one and five pupils participated, re-
spectively. The rest did not wish to take part or parental
permission was not given. In the other four classes, al-
most all pupils participated. In the different measures
the number of participants varied between 65 and 69,
because some children were absent from some testing
A subgroup of 26 children (16 girls and 10 boys),
comprising all children with imaginary companions
from four classes, were given a more thorough ques-
tionnaire and interviewed about their companions. In
the second wave of data collection, when an imaginary
companion questionnaire was given to the children,
two classes from the original sample dropped out be-
cause their teachers could not spare that extra time.
The Activity Questionnaire (AQ). The AQ
(Hoff, 2000) is a measure of involvement in creative
activities and hobbies and it includes a question about
whether the children have or have once had imaginary
companions. The questionnaire also concerns whether
the children engage in any creative hobbies (e.g., draw-
ing and writing stories), whether they spend a great
deal of time fantasizing, and whether they remember
their dreams. The children also indicate whether they
have invented their own games or built their own toys.
The maximum score was 10 (excluding the question
about imaginary companion from the statistical analy-
sis). The questions about creative activities and hob-
bies were shown to be related to the Creative Func-
tioning Test (Smith & Carlsson, 1990, 2001) and the
questionnaire in its present form with the Unusual
Uses Test (r= .34, p= .05 in Hoff, 2000; rs= .44, p=
.01 in Hoff & Carlsson, 2002).
170 Creativity Research Journal
E. V. Hoff
The Creative Functioning Test (CFT). The
Creative Functioning Test (Smith & Carlsson, 1990,
2001) is a measure of cognitive flexibility, which is
the ability to shift flexibly between imaginative and
rational thought. In CFT, a picture stimulus depicting
a black-and-white still life of a bottle and a bowl is
shown in repeated rapid presentations on a computer.
Shadings and diffuse contours build up the picture,
making it fairly easy to imagine other contents (e.g., a
body or a landscape). To begin with, in the increasing
series, the stimulus is presented for a very short time
(0.02 sec), and for every other presentation the expo-
sure time is prolonged until the participant describes
the picture content correctly (the longest possible ex-
posure time is 3.6 sec). Along the way to perceiving
the content objectively, a number of subjective inter-
pretations are often reported. When the participant
has perceived the actual content of the picture, the
procedure is reversed. In the decreasing series, the
picture is presented at shorter and shorter exposure
times and the session is finished when the stimulus
can no longer be discerned. The participants are in-
structed that pictures will be shown very briefly (but
not that it is the same picture). The participants are
told to describe what they see on the screen, even if
they are not quite certain.
The scoring in this study focuses on the decreasing
series, where new interpretations or recollections of
subjective themes from the increasing series are regis-
tered. This dimension captures an ability to shift from
rational (objective) thought to more imaginative (sub-
jective) cognition, an ability closely related to creativ-
ity (Smith & Carlsson, 1990). Objective perception
supposedly affects the viewer such that when correct
recognition has been attained, a low creative person
will inhibit subjective interpretations during the de-
creasing series. On the other hand, a highly creative in-
dividual will be able to abandon rational thought and
let the subjective representational world influence her
or his perception to a considerable extent. The scoring
involves six levels of creativity defined in the manual
(Smith & Carlsson, 1990, 2001).
The author and another judge scored the CFT proto-
cols independently, and in cases of disagreement a
third judge also made an assessment. According to the
manual (Smith & Carlsson, 2001), the retest correla-
tion for a group of children was .71. CFT has been vali-
dated through correlation with other measures, for ex-
ample, researchers rated by independent raters on
originality and richness of ideas (G= .67), preschool
children’s drawings rated by professional artists (G=
The Unusual Uses Test (UUT). The Unusual
Uses Test (Guilford, 1967) is regarded as a measure of
fluency of ideas. In the UUT, the participants make up
as many alternative uses as possible for a well-known
object, for example a newspaper or a brick. In this
study, the Unusual Uses Test was adapted to function
as a test for children. Empty milk packages were con-
sidered to be well-known objects for children and
therefore suitable for this study. The children were
asked to write down as many uses they could think of in
15 min. Regarding scoring, the total number of uses
was counted and every suggestion was given 1 point.
The self-image inventory: How I Think I Am. How
I Think I Am is a Swedish self-image inventory for
children (Ouvinen-Birgerstam, 1985/1999) and con-
sists of five subscales. The subscales measure (a) skills
and abilities—for example, “I’m good at math,” “Other
people do things better than I do”; (b) physical self-im-
age, health, and appearances—for example, “I don’t
care about my looks,” “I often feel clumsy”; (c) mental
well-being—for example, “I easily get angry,” “I’m a
happy person”; (d) relationship with parents—for ex-
ample, “My parents trust me,” “In my family, we fight
a lot”; and (e) relationships with peers and others—“I
have many friends,” “I feel different from others.
There are 72 items in the inventory. Each item has
four response alternatives, “Agree completely,” “pretty
much agree,” “hardly agree at all,” “disagree com-
pletely,” which were scored +2, +1, –1 and –2, respec-
tively, or conversely, depending on whether the ques-
tion dealt with a positive or negative matter. The
greater the sum, the better the self-image. Maximum
score is 144 and minimum –144.
According to the manual (Ouvinen-Birgerstam,
1985/1999), inter-item agreement showed alphas of
.91 through .93, and retesting a correlation of .74. The
inventory has also been validated through comparison
with an adjective list (r= .75).
The questionnaire about imaginary companions.
A questionnaire (see Appendix) about imaginary
companions provided further information about these
children and their companions. The information con-
cerned, for example, appearances (shape, size, age),
Creativity Research Journal 171
Imaginary Companions, Self, and Creativity
name, and sex (and other characteristics) of the com-
panions, the way they interacted (whether they played
in a fantasy world or with real toys), and whether they
engaged real peers in the play with their imaginary
The questionnaire items represent participants
found to be meaningful for children with imaginary
companions; the participants were identified through a
pilot study with unstructured interviews. The pilot
study participants were fourth graders—seven girls
and five boys. They were interviewed between 30 and
60 min during school hours about their imaginary com-
panions. However, some questionnaire items did not
provide any interesting information. For example,
number 13, concerning age of the friend, were ex-
cluded from the presentation of the results, as there was
very little variation. Eighty-eight percent of the chil-
dren had companions of the same age. Other less quan-
tifiable questions—for example question number 1,
about the name of the companion, and number 22,
about the content of the play—have been dealt with in
an article about the forms and functions of imaginary
companions (Hoff, 2005).
The interview. The children were also inter-
viewed, though much of this material is not relevant to
this article. However, one interview variable was
deemed relevant as well as useable in the quantitative
statistical analysis, namely that concerning the inde-
pendence of the make-believe friend—termed charac-
ter depth in the results section. Recall here that imagi-
nary companions may be associated with creativity
because they provide children with an opportunity to
cognitively elaborate. In the interview, the children
were asked about their roles and the roles of their pre-
tend playmates. There were questions concerning the
following: Who decided when they were about to
meet? Who comforted whom? Who urged whom?
Who made up mischief or suggested adventure? Who
stopped the play when things had gone too far? Who
determined what was right and wrong? There was also
a question about whether the children had learned any-
thing from their imaginary companion or whether the
companions had helped the children grow as individu-
als. To be assigned to the category “deep character” at
least two affirmative answers were required in re-
sponse to the questions of whether the child, during the
interview, had expressed that the playmate had showed
independent behavior or intentions (e.g., having taught
the child something or encouraged the child to be mis-
chievous) and whether the companion was described
as having independent characteristics compared with
the child (e.g., having its own relatives or that an elabo-
rated imaginary world was linked to the play). Besides
the author, a co-rater also categorized this variable.
The test administrator (the author) visited the chil-
dren during school hours. The first time the test admin-
istrator met the children, she presented the project, re-
lated what kind of questions would be posed to the
children, and asked them to return the parental permis-
sion form. On the next visit, four of the measures were
taken in the classroom. Within a few weeks’ time, the
last creativity measure (CFT) was taken individually in
a separate room. On the last visit, the imaginary com-
panion questionnaire was filled out by the subgroup of
26 children and the interviews were made.
The total mean on the AQ was 5.0 (SD = 2.6, range:
0–11). Inter-item agreement gave a Cronbach’s <α> of
.64. There was no gender difference on AQ (Table 1).
From the AQ it was concluded that 36 of the 69 chil-
dren (52%) currently had or had at one time had imagi-
nary companions—25 of the girls (71%) and 11 of the
boys (32 %). There was a significant gender difference
(Fisher’s Exact Test, p= .01).
The six levels of creativity defined in the manual
(Smith & Carlsson, 1990, 2001) were collapsed into
three larger groups. The total mean for the CFT scores
was 2.1 (SD = 0.8, range: 1–3). The interrater reliabil-
ity was 0.84 (Kappa). No gender difference was found
(Table 1).
For the Unusual Uses Test, the mean number of pro-
posed uses was 4.6 (SD = 2.8, range: 0–12). No signifi-
cant gender difference was found (Table 1).
172 Creativity Research Journal
E. V. Hoff
2The creativity and self-image results have been presented in an
earlier article (Hoff & Carlsson, 2002), where the relation between
these variables was the main interest. The relation between these
variables and imaginary companions has not been presented earlier.
The overall mean of How I think I Am was 77.2 (SD
= 31, range: –14 through 140) compared to the refer-
ence group mean of 68 (Ouvinen-Birgerstam, 1999).
There was no gender difference in total self-image.
The only subscale that showed a gender difference was
physical self-image, t(67) = –2.1, p= .04, to the advan-
tage of boys. Boys’mean was 17.1 (6.7) and girls’ 13.5
(7.4; see Table 1). In comparison with the reference
group (the reference mean of skills and abilities was
9.8; physical self-image, 12.8; mental health, 13.5;
parent relationships, 19; and relationships with friends
and others, 12.7), the participants of this study scored
somewhat higher in all subscales except on the mental
well-being scale, where they had average scores.
Imaginary Companions and Creativity
The children with imaginary companions scored
significantly higher on two of the creativity measures:
the AQ (U= 335, 5, p< .01) and The Unusual Uses Test
(U= 375.5, p< .01) with Mann Whitney U. The Cre-
ative Functioning Test showed a tendency in the same
direction (see Table 2).
Imaginary Companions and Self-Image
There were some significant self-image differences
between the children with imaginary companions and
those without. Those who had pretend playmates scored
Creativity Research Journal 173
Imaginary Companions, Self, and Creativity
Table 1. Descriptives of the Creativity Measures and the Self-Image Inventory
How I Think I Am
Participants AQ CFT UUT
Skills &
Relations Total
Girls (n= 35)
M5.6 2.0 5.0 13.3 13.5* 13.0 20.2 14.7 74.8
SD 2.8 0.9 2.9 7.8 7.4 8.4 7.5 7.4 28.6
Boys (n= 34)
M4.4 2.2 4.2 12.4 17.1* 14.3 21.5 14.4 79.8
SD 2.1 0.8 2.7 9.9 6.8 10.6 6.8 8.0 33.5
Total (N= 69)
M5.0 2.1 4.6 12.9 15.3 13.7 20.8 14.6 77.2
SD 2.6 0.8 2.8 8.8 7.2 9.5 7.2 7.6 31.0
Note. AQ = Activity Questionnaire; CFT = Creative Functioning Test; UUT = Unusual Uses Test.
*With an alpha level of .05, there was a difference between girls and boys.
Table 2. Imaginary Companions (i.c.) in Relation to the Creativity Measures and Self-Image Inventory
Had Never Had i.c. (n = 33) Had/or Had Once Had i.c. (n = 36)
M. Whitney
Uor TTest
Measures MSDMRank MSDMRank U/t p
Creativity U
AQ 3.9 1.8 27.2 5.9 2.8 42.2 335.5 .01
CFT 2.0 0.9 30.9 2.2 0.8 35.8 463 .28
UUT 3.6 1.9 28.2 5.4 3.2 40.1 375.5 .01
Self-image t
Skills and abilities 14.4 7.7 11.5 9.7 1.4 .17
Physical self 16.2 8.2 14.5 6.3 1.0 .34
Mental health 16.7 7.7 10.9 10.2 2.7 .01
Parent relations 21.8 6.4 20.0 7.8 1.0 .31
Peer relations 17.0 6.1 12.3 8.3 2.7 .01
Total self-image 86.1 27.6 69.1 32.1 2.3 .02
Note. AQ = Activity Questionnaire; CFT = Creative Functioning Test; UUT = Unusual Uses Test.
loweronthesubscaleofmentalwell-being, t(67) = 2.7, p
= .01, and on the subscale regarding relationships with
friends and others, for example teachers, t(67) = 2.7, p=
.01. The overall self-image was also significantly lower
for those with make-believe friends, t(67) = 2.3, p= .02.
Table 3 demonstrates the results. Compared with the ref-
erence material from the manual (Ouvinen-Birgerstam,
1999), the children with imaginary companions scored
lower than average on mental well-being. Their mean
was 10.9 and the mean of the reference group 13.5. On
relationships to others and overall self-image, the chil-
dren with imaginary companions had means equal to
those in the reference data.
The Subgroup With Imaginary Companions
There were no only children in the subgroup of 26
children for whom sibling data were obtained. Seven
children had one sibling, 15 children had two siblings,
and four children had three siblings. There were 10
first-born children among the sample and 16 middle or
Questionnaire about imaginary companions.
Most children had their make-believe friends (kept old
ones or acquired new) after the age of 7. The reason for
this age division was that Swedish children started
school at the age of 7 and it was deemed interesting to
see whether this influenced the imaginary companion-
ship. Ten children had their make-believe friends be-
fore age 7, 11 children kept old ones after school start,
and five acquired new ones. A majority of the children
only had one imaginary companion. The number of
children who played in an imaginary world with their
pretend playmates almost equaled the number who
played with their real toys. Most children played alone
with their companions. The make-believe friends most
often appeared in the shape of same-sexed humans.
There were very few differences between girls’ and
boys’ companions. The only significant finding was
that girls were more likely to play with their imaginary
companions together with a real playmate than were
boys (Fisher’s Exact, p= .04). Only one of the boys re-
ported playing with real friends while playing with his
imaginary companions. There was also a tendency for
girls to play more in paracosms than for boys to do so
(Fisher’s Exact, p= .10). The frequencies are presented
in Table 3.
Character depth of the imaginary companions.
The make-believe friends varied considerably re-
garding the level of elaboration of character. Their
characters were sometimes depicted as very much
alive and with an independent will, and sometimes the
characters were copies of the children who had in-
vented them. I will demonstrate some of these differ-
ences with quotations from the interviews. For the
quantitative analysis, the characters of the imaginary
companions were divided into the categories deep and
shallow. Apart from the author, a co-assessor also
made the categorization. The interrater reliability was
.77 (Kappa).
Deep characters. Some of the children gave ex-
amples of situations when their imaginary companions
were so much alive that the children almost appeared to
174 Creativity Research Journal
E. V. Hoff
Table 3. Aspects of Imaginary Companions (i.c.) From the
Aspects of i.c.
(n= 16)
(n = 10)
(N = 26)
Age when having i.c.
Before 7 6 5 11
After 7a10 5 15
Number of i.c.
One i.c. 8 8 16
More than one i.c. 8 2 10
Play context
Play in real world 11 3 14
Play in paracosm 5 7 12
Character depth
Shallow character 7 5 12
Deep character 9 5 14
Number of places
Plays at home 7 7 14
Different places 9 3 12
Type of play
Plays alone 7 9 16
Plays with others 9 1 10
Sex of i.c.
Same-sex i.c. 11 9 20
Opposite-sex i.c. 5 1 6
Type of i.c.
Human 10 7 17
Animal 6 3 9
aComprising both those who acquired their imaginary companions
before 7 and kept them until after 7 years of age and those who ac-
quired imaginary companions after the age of 7.
have forgotten that they were their own imaginary con-
Frida allowed her make-believe friend, “Nicki,” to
(secretly) join her when she played with real friends.
But sometimes she sent Nicki away when she was not
available to play with her, out of consideration that
Nicki might become bored: “She’s usually allowed to
be there, but sometimes she has to go too. I mean, I
can’t, I just can’t … it would be boring for her if I
couldn’t speak to her.” Harriet reported that, in the be-
ginning, she felt timid in the presence of the imaginary
companion: “I used to be pretty shy of her.
In a few cases, the make-believe friends had
well-considered characters. When asked what he knew
about his friends, Elvis described them in the following
Well they are … from China. They are … live on
being happy and they eat too, but they have to be
happy in order to live. Then they are mostly kind,
to me. But they tease each other. Especially Pep-
per /the companion’s name/.
Ida reported what her companion taught her:
Interviewer: Can “Knubbis”/Chubby/ tell you things
you don’t know? Can he teach you
Ida: Yes, he can teach me about where he was
before, when he didn’t know who I was
Interviewer: Aha, where was he before then?
Ida: He was on a planet, and I think it was
Mars or something.
Interviewer: Aha.
Ida: Or Jupiter, some planet anyway and then
… he was up there and then there was a, I
think there was a fire there, that’s why …
he left. And then he fell down from the
planet to earth. And then he came … and
then I just saw him.
Shallow characters. On the other hand, several
children did not know much about their pretend play-
mates. They appeared when the children were bored or
sad and kept them company or comforted them as long
as the children needed them and then they vanished.
These less elaborated companions looked like their in-
ventors and did not have much will of their own.
Aron’s make-believe friend was not clearly distin-
guished from Aron himself and when they were differ-
ent, Aron was the active part and “Kalle” the passive. A
trick he and Kalle used to play on his parents was to
hide things, and Aron related that he was the one who
made up the trick. On other occasions, their wills were
equal. To the question about who urged the other when
they were going out on an adventure, he responded:
“Nobody, because we both dared.
Another similar example was Hilda and her pretend
playmate “Madeleine.” They were very much alike.
She told of a kind of mischief she and Madeleine used
to get into:
Interviewer: Who invented the mischief? Was it you
or she?
Hilda: It was none of us. We used to make them
up together. /…/
Interviewer: Was it ever like the pretend playmate
was with you and said: “Let’s do this”
and that you said: “No, I don’t want to”?
Hilda: No.
Interviewer: You always had the same opinion?
Hilda: Yes.
Interviewer: So it never happened that you had to en-
courage the pretend playmate to join you
in something she didn’t want to do?
Could you have different opinions?
Hilda: No, mostly not.
Interviewer: No, is there an example of a situation
when you had different opinions?
Hilda: She always wanted to be with me and I
really wanted to be with her and all.
Imaginary companions and creativity in the sub-
group. Among the 26 children in the subgroup with
imaginary companions, there were some relationships
with the creativity measures. The UUT and the AQ
were selected as creativity measurements for this
smaller sample, as the CFT dimension did not correlate
significantly with the whole sample. Means, mean
ranks, and statistical figures are presented in Table 4.
Some aspects of the imaginary companionship were
related to higher creativity scores in the children.
Those children who had (kept old ones or acquired
new) imaginary companions after the age of 7 (U=
43.5, p= .04) had a greater number of make-believe
Creativity Research Journal 175
Imaginary Companions, Self, and Creativity
friends (U= 23.5, p= .001), had companions with
more elaborated independent characters (U= 34.5, p=
.01), played at different places to a greater extent (U=
20.5, p= .001), and played more often together with
real playmates; their make-believe friends at the same
time (U= 30.5, p= .04) scored significantly higher on
the AQ. There was also a tendency indicating that
paracosmic play was more creative (p= .12). The rela-
tionships were weaker with the Unusual Uses Test. The
only variable showing significantly separated creativ-
ity scores was the number of imaginary companions (U
= 40.5, p= .04). Tendencies appeared for character
depth, number of locations for the play, and combined
involvement with real friends and for pretend play-
mates of the opposite sex.
Imaginary companion and self-image in the sub-
group. Within the subgroup, there were no signifi-
cant differences in self-image scores across the differ-
ent aspects of the companionship.
Half of the fourth graders in this study reported hav-
ing imaginary companions at the time of the interview
or having had such companions earlier in their life.
One of the main findings was that having imaginary
companions was related to two out of three estimates
of creative potential. The third creativity measure
showed a tendency in the same direction. Few previous
studies have focused on 10-year-olds, though some
have been able to show associations between creativity
and imaginary companions (Schaefer, 1969;
Seiffge-Krenke, 1997; Singer, 1961). On the other
hand, there are quite a few case studies discussing cre-
ative adult people who had imaginary companions as
children (Myers, 1979; M. Taylor, 1999).
Regarding self-image, the children with imaginary
companions, as compared to those without, described
themselves to a greater extent as having few friends,
being different from others, and having lower psycho-
176 Creativity Research Journal
E. V. Hoff
Table 4. Creativity Measures and Different Aspects of Imaginary Companions (i.c.) in Subgroupa
Activity Questionnaire Unusual Uses Test
Aspects of i.c. MSDMRank UpMSDMRank Up
Age when having i.c.
Before 7 5.6 2.1 10.0 43.5 .04 4.1 2.3 12.4 71.0 .55
After 7b7.3 2.4 16.1 5.3 4.2 14.3
Number of i.c.
One i.c. 5.6 2.2 10.0 23.5 .001 3.6 2.6 11.0 40.5 .04
More than one i.c. 8.2 1.8 19.2 6.8 4.8 17.4
Play context
Play in real world 5.8 2.7 11.0 54.0 .12 4.5 3.9 12.2 68.0 .40
Play in paracosm 7.3 2.0 15.6 5.1 3.3 14.6
Character depth
Shallow character 5.3 2.4 9.4 34.5 .01 3.6 3.3 10.5 47.5 .06
Deep character 7.8 1.8 17.0 5.9 3.5 16.1
Number of places
Plays at home 5.3 2.2 9.0 20.5 .001 4.0 3.5 11.4 54.5 .12
Different places 8.2 1.5 18.8 5.7 3.5 16.0
Type of play
Plays alone 5.9 2.5 10.4 30.5 .04 4.2 3.5 11.1 42.0 .15
Plays with others 7.9 1.5 18.4 5.9 2.5 17.3
Sex of i.c.
Same-sex i.c. 6.3 2.4 12.6 48.5 .30 4.0 3.1 12.0 38.5 .10
Opposite-sex i.c. 7.4 2.2 16.1 6.8 4.1 17.5
Type of i.c.
Human 6.6 2.2 14.6 67.0 .61 4.1 2.9 16.1 53.0 .20
Animal 6.8 2.9 12.9 6.1 4.3 12.1
an= 26. bComprising both those who acquired their imaginary companions before 7 and kept them until after 7 years of age and those who ac-
quired imaginary companions after the age of 7.
logical well-being. They also had lower self-image
scores on the overall scale. However, it is important
that we do not, based on these results, ascribe severe
psychological problems to the inventors of imaginary
companions. Compared with the reference data in the
manual of How I Think I Am (Ouvinen-Birgerstam,
1985/1999), the mean score for children with pretend
playmates is noticeably discrepant for mental self-im-
age, but about average for relationships with others and
overall self-image. The sample in this study as a whole
was well above average, probably due to, among other
things, a larger percentage of children from mid-
dle-class homes. Other studies have also shown the dis-
advantages of children with make-believe friends
(Bouldin & Pratt, 1999, 2002; Harter & Chao, 1992).
Bouldin and Pratt (2002) showed that anxiety levels
(judged from parental reports) were higher in 3- to
8-year-old children with imaginary companions com-
pared with those who did not have such companions.
Harter and Chao (1992) demonstrated that 3- to
6-year-old children with make-believe friends, as
judged from preschool teacher reports, were less com-
petent and less socially accepted by peers.
On the other hand, still other studies have demon-
strated the advantages of the possessors of imaginary
companions and shown that these children are superior
in some respects to other children (Manosevitz et al.,
1973; Singer & Singer, 1992).
Although this study and some others have demon-
strated that children with imaginary companions might
be socially less capable, it can be fruitful to speculate
on how these children would have coped had they not
had these companions at all. In another report on the
functions of imaginary companions (Hoff, 2005), it
was shown that, according to the children, imaginary
companions not only had a social compensatory func-
tion, but also provided social practice and enhanced the
children’s social competence. Other researchers have
also argued in line with these results. Gleason (2002)
maintained that make-believe friends may provide
practice in conceptualizing relationships. Harter and
Chao (1992) purported that the competence of the
child in general might be increased through the inven-
tion of an imaginary companion.
In the field of social cognition, the importance of il-
lusions in maintaining self-esteem has been put for-
ward (S. Taylor & Brown, 1999). In this context, hav-
ing an imaginary companion can be a way of
maintaining an illusion of social competence.
More complex peer interaction is developed in the
early school years and social competence is an increas-
ingly important skill (Erikson, 1950/1993; Piaget,
1968). The imaginary companion might function as a
transitional phase for some children, allowing them to
gradually develop the skills for managing the role-tak-
ing and rule-oriented interactions that real friends de-
mand (Sugarman & Jaffee, 1989).
However, it is still unclear whether the children are
assisted or inhibited by their imaginativeness as con-
cerns their social and mental functioning. Longitudinal
research, in which socially isolated children both with
and without imaginary companions participate, is
needed to answer the question of whether the invention
of pretend playmates increases or at least maintains
children’s isolation, or whether this kind of fantasy
play actually provides social practice during a transi-
tional phase and helps the possessors of imaginary
companions to become more mentally healthy and so-
cially capable in the long run.
The high frequency of make-believe friends (52%)
reported in this study may seem surprising. However,
compared to some other studies, the present incidence
is actually somewhat low (e.g., 65% in Singer &
Singer, 1992, and 63% in M. Taylor, 1999). One reason
for the somewhat lower frequency might be that some
of the 10-year-olds had actually forgotten their early
pretend playmates. An important question within this
research area is whether the parents or the children
themselves should constitute the source of information
when investigating imaginary companions. Severe
problems with the use of parental reports have been
disclosed, especially for older children, as parents are
seldom aware of their make-believe friends (M. Taylor,
1999). A combination of accounts from parents and
children would in most cases be optimal. In this study
only children’s reports were utilized. Another explana-
tion for the somewhat lower occurrence rate might be
that the child participants were not explicitly informed
that anthropomorphized toys could be counted as
imaginary companions. In this study, only three chil-
dren had real-world props in the shape of stuffed ani-
mals. Clearer instructions and parental reports, biased
or not, may have identified more anthropomorphized
toys and some of the forgotten early instances of imagi-
nary companions.
Ten-year-old children have been demonstrated to be
well-functioning participants as well as reliable infor-
mants (Andersson, 1998; Garbarino & Stott, 1992),
Creativity Research Journal 177
Imaginary Companions, Self, and Creativity
perhaps more reliable than younger children who
sometimes have difficulties in separating reality and
imagination (M. Taylor, 1999). In one respect,
10-year-olds are also more trustworthy than teenagers,
because they are still engaged in pretend play or have
fairly recently abandoned this kind of play, whereas
teenagers’ experiences are more distant.
The Subgroup With Imaginary Companions
In this study, no support was obtained for the as-
sumption that only and first-born children tend to in-
vent imaginary companions more often. In the smaller
sample of 26 children for which family data were col-
lected, there were no only children and just 10 of 26
were first born. However, no statistical evidence could
be produced, as there were no family data collected
among the children who did not have imaginary com-
In the subgroup, more detailed information about
the characteristics of make-believe friends was used to
uncover particularly creative features of imaginary
companions. A greater number of pretend playmates, a
more elaborated independent character of the friend
and greater variation in places where the companions
appeared, indicated a more creative inventor. Engaging
other real friends in the pretend play also implied
greater creativity. There was a tendency for the
paracosmic imaginary play to be more creative. To
summarize, the more variation and elaboration in the
imaginary play, the more creative the child. These re-
sults must be replicated in a larger sample, and to ac-
complish this the questionnaire about imaginary com-
panions will need some revision. For example,
questions (in this study included in the interview) re-
garding the independence of the imaginary companion
should be added. The result concerning character depth
needs to be regarded with great caution. The children’s
memory capacity might have been a confounding fac-
tor, as children who had make-believe friends after
school entrance were also assessed to be more creative.
The children with elaborated companions might have
remembered more details concerning their pretend
playmates compared to those children who had had
companions in earlier years.
Regarding gender differences, girls, as expected,
more often had imaginary companions. Other, previ-
ously demonstrated gender differences were not cor-
roborated, which might be due to the small subsample.
Some of the expected differences concerned whether
the companion had the shape of a human or an animal
and whether the sex of the pretend playmate was the
same as or opposite the child’s. In this study, only 42%
of the boys had imaginary animal friends, whereas
60% of the girls did, although the opposite proportions
were expected. Regarding sex of the companions, the
participants showed the anticipated differences,
though they were nonsignificant: 45% of the girls had
companions of the opposite sex, whereas only 11% of
the boys had a female companion. Despite the small
number of participants, one new finding was detected:
The girls in this study, as compared to the boys, re-
ported playing more often with other real playmates to-
gether with their pretend playmates. This finding needs
to be verified in further studies. One speculation about
the reason for the few gender differences in this study
is that the more varied gender roles in Sweden might
cause Swedish boys and girls to invent more similar
friends than do children in other countries where the
gender roles are more traditional (Boski, 2002). Future
studies should address cultural comparisons regarding
possible gender differences in imaginary companions.
Ames, L. B., & Learned, J. (1946). Imaginary companions and re-
lated phenomena. The Journal of Genetic Psychology, 69,
Andersson, G. (1998). Barnintervju som forskningsmetod [Inter-
views with children as research method]. Nordisk Psykologi,
50, 18–41.
Bach, S. (1971). Notes on some imaginary companions. Psychoana-
lytic Study of the Child, 26, 159–171.
Bender, L., & Vogel, B. F. (1941). Imaginary companions of chil-
dren. American Journal of Orthopsychiatry, 11, 56–66.
Boski, P. (2002, Sept). A seminar on masculinity—femininity. The
seminar conducted at the Workshop in Cultural Psychol-
ogy/Methodological Issues, Lund, Sweden.
Bouldin, P., & Pratt, C. (1999). Characteristics of preschool and
school-age children with imaginary companions. Journal of
Genetic Psychology, 160, 397–410.
Bouldin, P., & Pratt, C. (2002). A systematic assessment of the spe-
cific fears, anxiety level, and temperament of children with
imaginary companions. Australian Journal of Psychology, 54,
Cohen, D., & Mackeith, S. (1991). The development of imagination:
The private worlds of childhood. New York: Routledge.
Erikson, E. H. (1993). Barnet och samhället [Childhood and soci-
ety]. (J. Rössel, Trans.). Stockholm: Natur och kultur. (Original
work published 1950)
178 Creativity Research Journal
E. V. Hoff
Garbarino, J., & Stott, F. M. (1992). What children can tell us:
Eliciting, interpreting, and evaluating critical information from
children. San Francisco: Jossey-Bass.
Gleason, T. R. (2002). Social provisions of real and imaginary rela-
tionships in early childhood. Developmental Psychology, 38,
Gleason, T. R., Sebanc, A. M., & Hartup, W. W. (2000). Imaginary
companions of preschool children. Developmental Psychology,
36, 419–428.
Guilford, J. P. (1967). The nature of human intelligence. New York:
Harter, S., & Chao, C. (1992). The role of competence in children’s
creation of imaginary friends. Merrill-Palmer Quarterly, 38,
Hoff, E. (2000). En målares gåta: En berättelse som utgångspunkt för
att studera kreativitet hos 10-åringar [A painter’s mystery: A
story as starting-point to study creativity in 10-year-old chil-
dren]. Nordisk Psykologi, 52, 37–77.
Hoff, E. V. (2005). A friend living inside me: The forms and func-
tions of imaginary companions. Imagination, Cognition and
Personality, 24, 151–189.
Hoff, E. V., & Carlsson, I. (2002). Shining lights or lone wolves?
Creativity and self-image in primary school children. Journal of
Creative Behavior, 36, 17–41.
Hurlock, E. B., & Burstein, M. (1932). The imaginary playmate: A
questionnaire study. Journal of Genetic Psychology, 41,
Mackeith, S. A. (1982). Paracosms and the development of fantasy
in childhood. Imagination, Cognition and Personality, 2,
Manosevitz, M., Fling, S., & Prentice, N. M. (1977). Imaginary com-
panions in young children: Relationships with intelligence, cre-
ativity and waiting ability. Journal of Child Psychology and
Psychiatry, 18, 73–78.
Manosevitz, M., Prentice, N. M., & Wilson, F. (1973). Individual and
family correlates of imaginary companions in preschool chil-
dren. Developmental Psychology, 8, 72–79.
Martindale, C. (1989). Personality, situation, and creativity. In J. A.
Glover & R. R. Ronning (Eds.), Handbook of creativity. Lon-
don: Plenum.
Myers, W. A. (1979) Imaginary companions in childhood and adult
creativity. Psychoanalytic Quarterly, 48, 292–307.
Nagera, H. (1969). The imaginary companion: Its significance for
ego development and conflict solution. Psychoanalytic Study of
the Child, 24, 165–196.
Ouvinen-Birgerstam, P. (1999). Jag tycker jag är [How I think I am].
Stockholm: Psykologiförlaget. (Original work published 1985)
Piaget, J. (1968). Six psychological studies. (A. Tenzer, Trans.). New
York: Vintage Books.
Schaefer, C. E. (1969). Imaginary companions and creative adoles-
cents. Developmental Psychology, 1, 747–749.
Seiffge-Krenke, I. (1997). Imaginary companions in adolescence:
Sign of a deficient or positive development? Journal of Adoles-
cence, 20, 137–154.
Seiffge-Krenke, I. (2000). Ein sehr spezieller Freund: Der
Imaginaere Gefaehrte [A very special friend: The imaginary
companion]. Praxis der Kinderpsychologie und
Kinderpsychiatrie, 49, 689–702.
Singer, D. G., & Singer, J. L. (1992). The house of make-believe:
Children’s play and the developing imagination. Cambridge,
MA: Harvard University Press.
Singer, J. L. (1961). Imagination and waiting ability in young chil-
dren. Journal of Personality, 29, 396–413.
Smith, G. J. W., & Carlsson, I. M. (1990). The creative process: A
functional model based on empirical studies from early child-
hood to middle age. Madison, CT: International Universities
Smith, G. J. W., & Carlsson, I. M. (2001). CFT: Creative functioning
test. Lund, Sweden: Department of psychology. (Original work
published 1990)
Somers, J. U., & Yawkey, T. D. (1984). Imaginary companions: Con-
tributions of creative and intellectual abilities of young chil-
dren. Journal of Creative Behavior, 18, 77–89.
Sperling, O. E. (1954). An imaginary companion, representing a
prestage of the superego. Psychoanalytic Study of the Child, 9,
Sugarman, A., & Jaffee, L. S. (1989). A Developmental Line of
Transitional Phenomenon. I. In M. G. Fromm & B. L. Smith
(Eds.), The facilitating environment: Clinical applications of
Winnicott’s theory (pp. 81–129). Madison, CT: International
University Press.
Svendsen, M. (1934). Children’s imaginary companions. Archives of
Neurology and Psychiatry, 32, 985–999.
Taylor, M. (1999). Imaginary companions and the children who cre-
ate them. New York: Oxford University Press.
Taylor, M., & Carlson, S. M. (1997). The relation between individual
differences in fantasy and theory of mind. Child Development,
68, 436–455.
Taylor, S. E., & Brown, J. D. (1999). Illusion and well-being: A so-
cial psychological perspective on mental health. In R. F.
Baumester (Ed.), The self in social psychology (pp. 43–67).
Philadelphia: Psychology.
Vostrovsky, C. (1895). A study of imaginary companions. Educa-
tion, 15, 393–398.
Creativity Research Journal 179
Imaginary Companions, Self, and Creativity
180 Creativity Research Journal
E. V. Hoff
... Experiences of childhood ICs are relatively common, but methodological and conceptual variations prevent precise incidence rates Accordingly, previous studies have reported figures as high as 65% (Armah & Landers-Potts, 2021), though most typically between 20% and 35% (Giménez-Dasi et al., 2014). A further feature is that IC character formation ranges from shallow (i.e., basically copies of the children who invented them) to deep (i.e., characters that seem alive and with independent wills) (e.g., Fernyhough et al., 2019;Hoff, 2005). Further to these deep forms, some of the children in Hoff (2005) reported ICs that seemed so much 'alive' that they forgot these 'friends' were imaginary. ...
... A further feature is that IC character formation ranges from shallow (i.e., basically copies of the children who invented them) to deep (i.e., characters that seem alive and with independent wills) (e.g., Fernyhough et al., 2019;Hoff, 2005). Further to these deep forms, some of the children in Hoff (2005) reported ICs that seemed so much 'alive' that they forgot these 'friends' were imaginary. Research on accounts of deep ICs suggest that these instances might be best conceptualized as '(entity) encounter experiences' (EEs) (Little et al., 2021). ...
... In this context, some children (and adults) might possess the ability to generate EEs in different sensory modes on demand, in naturalistic settings, and during apparent normal waking states. This view of ICs, and especially as related to their deep forms, suggests these are hallucination-like experiences with potentially adaptive value for the development of social cognition (Davis et al., 2011), creativity (Hoff, 2005), or inner speech, i.e., internal voices that combine conscious thoughts and unconscious beliefs and biases to help interpret and process questions, ideas, or experiences (Fernyhough et al., 2019). Indeed, there is a burgeoning literature on creativity in relation to hallucinatory, metachoric, and syncretic experiences (e.g., D'Anselmo et al., 2020;Fink et al., 2014;Green, 1990;Wilson & Barber, 1981)all of which positively correlate with thin mental boundary functioning . ...
This study investigated relationships between thin mental boundary functioning, creativity, imaginary companions (ICs), and anomalous ‘(entity) encounter experiences.’ A convenience sample of 389 respondents completed the Revised Transliminality Scale, Oxford-Liverpool Inventory of Feelings and Experiences, Creative Experiences Questionnaire, Survey of Strange Events, and a measure of Childhood Imaginary Companions. Competing testing with path analysis found that the best-fitting model was consistent with the causal chain of ‘Thin Boundaries (transliminality and schizotypy) → Creative Experiences → ICs → (Entity) Encounter Experiences.’ These results suggest that deep-types of ICs (i.e., showing apparent independent agency) are perhaps most accurately characterized as syncretic cognitions versus hallucination-like experiences. The authors examine these findings relative to study limitations, as well as discussing the need for future research that approaches ICs as a special mental state that can facilitate allied altered-anomalous experiences. In this context, this study furthered understanding of relationships between conscious states related to mental boundaries, childhood imaginary companions, creative experiences, and entity encounters.
... An imaginary companion or friend has been defined as "a very vivid imaginary character (person, animal) with which a child interacts during his/her play and daily activities" (Taylor, 1999, p. 28), or a character "that does not actually exist, although the child treats it as though it does" (Bouldin, 2006, p. 19). An imaginary companion can sometimes be entirely invisible, and sometimes take the form of a stuffed animal or doll (Taylor, 1999;Singer & Singer, 1990;Hoff, 2004Hoff, , 2005. The imaginary companions can vary in the length of time, and they usually, but not always, are of a friendly character. ...
... On average, girls talk out loud to their make-believe friends more than boys. Their imaginary companions were more likely to be human (Singer & Singer, 1990;Hoff, 2005), and their IFP more cooperative and social (Papastathopoulos & Kugiumutzakis, 2007), than those of boys, keeping in with the generally greater verbal development of girls. ...
... Also, IFP can be hidden from a family member. Hoff (2004Hoff ( , 2005 documents numerous imaginary companions of ten-year-olds that seem to exist primarily within their creators' minds and are not shared with others. This paper has focused mainly on the psychological data that showed explicit engagements of pretenders, and did not discuss instances of children's IFP that are unobservable to the researchers or parents. ...
Full-text available
Review of psychological data of how children engage in imaginary friend play (IFP) shows that it involves a lot of explicit embodied action and interaction with the surrounding people and environments. However, IFP is still seen as principally an individualistic activity, where in addition to those interactions, the actor has to mentally represent an absent entity in imagination in order to engage in IFP. This capacity is deemed necessary because the imaginary companion is absent or not real. This paper proposes a proof of concept argument that radical enactivism can account for complex imaginary phenomena as imaginary friend play. Enactivism proposes to think of IFP in a fundamentally different way, as an explicitly embodied and performative act, where one does not need to mentally represent absent entities. It reconceptualises imagination involved in IFP as strongly embodied, and proposes that play environments have present affordances for social and normative interactions that are re-enacted in IFP-there is no "absence" that needs to be mentally represented first. The paper argues that IFP is performed and enacted in the world without having to be represented in the mind first, which best captures the social and interactive nature of this form of play.
... Having an imaginary companion is one of the most salient features of the fantasy lives of young children (Fernyhough et al., 2007). Hoff (2005) points out that there is a correlation between having an imaginary companion and creativity, and the more variation and elaboration in imaginary play is also related to the level of creativity. Gleason et al. (2003) found that participants who reported having an imaginary companion during childhood scored higher on measures of imagination than participants who reported no such companion. ...
... It has also been suggested that there is a connection between having an imaginary companion and creativity and imagination (Gleason et al., 2003;Hoff, 2005;Root-Bernstein&Root-Bernstein, 2006). Pearson et al. (2001) and Taylor et al. (2004) put forth that girls have more imaginary companions than boys. ...
Full-text available
Imagination is necessary for creative ideas to emerge. The creative imagination can be developed by suitable education programs especially by drama programs with suitable activities. This article presents findings on whether the effect of drama on the creative imagination of children in different age groups differentiate or not. The experiment group of this research is comprised of 60 children (30 from the age group of 10, 30 from the age group of 13) from a regular primary school and the control group is comprised of 60 children (30 from the age group of 10, 30 from the age group of 13) from another primary school both with the equal socioeconomic background in order to avoid children effect each other. The drama program was implemented to the experiment group of both age groups separately as 48 hours in a week. However, this program was not implemented to the control group. The Kujawski Creative Imagination Test was used as the measurement instrument. For analysis of the data, the independent samples t test and paired samples t test were utilized. From the results of the creative imagination test applied before and after the drama, it was shown that a drama program has a positive effect on development of the creative imagination of children. When the effects of drama in different age groups is examined, the prepared drama program is more effective on 10 year old children than 13 year old children. ÖZ: Hayal gücü, yaratıcı fikirlerin ortaya çıkması için gereklidir. Uygun eğitim programları, özellikle de uygun etkinliklerin yer aldığı drama programları aracılığıyla yaratıcı hayal gücü geliĢtirilebilir. Bu makale, farklı yaĢ gruplarındaki çocukların yaratıcı hayal güçleri üzerinde dramanın etkisinin farklılaĢıp farklılaĢmadığını ortaya koymaktadır. Bu araĢtırmanın deney grubunu bir ilköğretim okulundan 60 çocuk (10 yaĢ grubundan 30, 13 yaĢ grubundan 30) oluĢtururken, kontrol grubunu çocukların birbirinden etkilenmemeleri için sosyo-ekonomik olarak eĢ düĢünülen baĢka bir ilköğretim okulundan 60 çocuk (10 yaĢ grubundan 30, 13 yaĢ grubundan 30) oluĢturmaktadır. Drama programı, her iki yaĢ grubundaki deney grubuna ayrı ayrı haftada dört saat toplam 48 saat uygulanmıĢ, kontrol grubuna uygulanmamıĢtır. Ölçme aracı olarak Kujawski Yaratıcı Hayal Gücü Testi kullanılmıĢtır. Verilerin analizinde bağımsız örneklemli t testi ve eĢleĢtirilmiĢ gruplar için t testi kullanılmıĢtır. Drama öncesi ve sonrası uygulanan yaratıcı hayal gücü test sonuçlarına göre, drama programının çocukların yaratıcı hayal güçlerinin geliĢiminde olumlu bir etkisi vardır. Dramanın farklı yaĢ gruplarına göre etkisi incelendiğinde, hazırlanan drama programı 13 yaĢ çocuklarına göre 10 yaĢ çocuklarında daha fazla etkilidir.
... When weighing up potential courses of actions to the breach, we may entertain scenarios with our generalised others (Crossley, 2006) (Baldwin, Carrell & Lopez, 1990;Baldwin & Holmes, 1987). In the absence of opportunity to relate with others, or in loneliness, imaginary friendships are sometimes created as a compensatory strategy (Hoff, 2005) or to create opportunities for social practice and social competence (Gleason, 2002). New identities and ways of feeling may be formed through a process whereby the personal, collective and imagined evaluations of our selves are brought into consciousness and then assimilated; the more successful the integration between who we see ourselves as, and how we are seen by others, the more inclusive and integrated our self-concept becomes. ...
Research, theory and mental health policy draws attention to the importance of family, social networks, community, employers and learning contexts in maintaining mental health and inclusion. Yet the meaningful complexities of friendship to psychological health and public policy has not received sustained analysis, and policy emphasis is often restricted towards family relations. This study explores the friendship-experiences of seventeen people who have endured mental health difficulties, through a critical narrative inquiry of their stories of friendship. A hermeneutics of suspicion, involving stigma, feminist and mad studies is used to explore meaning within the narratives. The study reveals the participants’ stories of problems of daily living, illness and stigma, of friendship as freedom and recognition, and friendship’s contribution to personal agency and establishing a valued position in society. The study develops a perspective of how compassion in friendship has helped articulate and reframe identities to one’s self, to others, and to distress, and therefore the potential contribution of friendship to living with mental distress. The thesis argues that mental health studies have been dominated by institutionalised relationships, of which friendship has been made to fit into theoretical frameworks of family- and kin-relationships. The thesis presents an alternative view of friendship to aid in the reformulation of the varieties of social relationships shared by people through mental distress. Additionally, there have been very few narrative studies that explore the friendship experiences of people with mental ill health and this study adds to a growing literature.
... Developmental psychology research has shown that ICs may play an important role in children's cognitive [21] and verbal skills [146], as well as coping [36] and stress management [50]. ICs provide entertainment, playmates and even friendship that enable a child to overcome times of boredom and loneliness, and they can also be a source of comfort when the child is experiencing difficulties. ...
Full-text available
Maladaptive daydreaming (MD) is an excessive and vivid fantasy activity that interferes with individual's normal functioning and can result in severe distress. Research has shown that MD is a clinical condition associated with a number of personal, interpersonal and behavioral problems. Therefore, a need exists to differentiate MD from other mental activities that involve an excessive or otherwise problematic use of fantasy. These include, among others, daydreaming, mind wandering, dissociative absorption, fantasy proneness, sluggish cognitive tempo, lucid dreaming, and autistic fantasy. In this article, we examine the commonalities and differences between MD and these mental activities, to promote a better understanding of the MD phenomena and their specificity, and to foster the quality of its assessment in clinical settings. A clinical case study is employed to elucidate our analysis and to demonstrate the differential diagnosis of MD.
... Algumas pessoas também podem ler o amigo imaginário como um sinal de criatividade (Taylor, 1999;Hoff, 2005). E1 mencionou essa concepção em sua entrevista, quando comentou que sua prima que interagia com um amigo imaginário se tornou uma pessoa muito criativa e proativa. ...
Full-text available
A presente pesquisa buscou compreender como os profissionais da área de Psicologia e Psiquiatria, com atuação na clínica infantil, entendem a presença do fenômeno do amigo imaginário na infância. Três psicólogos e dois psiquiatras, com 5 anos ou mais de experiência, foram entrevistados sobre suas práticas profissionais. Os participantes foram recrutados em clínicas de atendimento à criança em uma cidade do interior do Rio Grande do Sul. A análise de conteúdo indicou que os profissionais consideravam o fenômeno um acontecimento saudável do desenvolvimento e facilmente diferenciável da sintomatologia da psicose. Além disso, acreditavam que a fantasia pode servir como um recurso para o profissional investigar a realidade psíquica da criança, facilitando a comunicação nos atendimentos. A partir desses resultados, buscou-se contribuir com a literatura nacional acerca do tema, ainda escassa, e com a atuação e conhecimento dos profissionais da área de saúde infantil e todos aqueles que estão em constante contato com crianças (profissionais escolares, familiares, etc.).
Full-text available
Review of psychological data of how children engage in imaginary friend play (IFP) shows that it involves a lot of explicit embodied action and interaction with surrounding people and environments. However, IFP is still seen as principally an individualistic activity, where, in addition to those interactions, the actor has to mentally represent an absent entity in imagination in order to engage in IFP. This capacity is deemed necessary because the imaginary companion is absent or not real. This article proposes a proof of concept argument that enactivism can account for complex imaginary phenomena as imaginary friend play. Enactivism proposes thinking of IFP in a fundamentally different way, as an explicitly embodied and performative act, where one does not need to mentally represent absent entities. It reconceptualizes imagination involved in IFP as strongly embodied, and proposes that play environments have present affordances for social and normative interactions that are reenacted in IFP—there is no “absence” that needs to be mentally represented first. This article argues that IFP is performed and enacted in the world without having to be represented in the mind first, which best captures the social and interactive nature of this form of play.
A curious childhood phenomenon that has received relatively little attention in developmental literature is the imaginary companion (IC). Increased recognition of the importance of imaginative play and a desire to stimulate children’s early cognitive development makes ICs a particularly relevant topic. The significant prevalence of ICs in the population has permitted a modest yet diverse range of research investigating the functions, correlates, and implications of ICs for the children that create them. This literature review summarizes some of this research in order to describe the functions and forms that ICs may take, as well as social and personality characteristics of children with ICs. It also examines the role that ICs may serve in cognitive and social development, particularly with respect to children’s acquisition of Theory of Mind. Finally, this article addresses ways to integrate ICs into other aspects of children’s lives, gaps in the existing literature, and potential directions for future research in the field.
Full-text available
Imaginary companions and the characteristics of the children who invent them were explored through interviews. Twenty-six children took part. The first aim was to explore the phenomenon in general, the second to investigate the functions of imaginary companions. General findings concerned the kinds of imaginary companions and some facts about the children who created them. The imaginary companions were mostly same-aged children, but some were different kinds of animal or fantasy creatures. The sources of inspiration varied, comprising friends and siblings. Having a place of their own, where the children could develop their imaginary play, was considered important. The study demonstrated various ways of conducting imaginary play. Imaginary companions functioned as inner mentors, assisting the children in their identity formation work. Imaginary companions were experienced by the children as giving comfort and company, bolstering self-regulation and motivation, enhancing their selves, expanding their personality potential, and finally, enriching their lives.
Conference Paper
The developmental significance of preschool children's imaginary companions was examined. Mothers of 78 children were interviewed about their children's social environments and imaginary companions (if their children had them). Results revealed differences between invisible companions and personified objects (e.g., stuffed animals or dolls) in terms of the pretend friends' stability and ubiquity, identity, and relationship with the child. Relationships with invisible companions were mostly described as sociable and friendly, whereas personified objects were usually nurtured. Mothers reported that personification of objects frequently occurred as a result of acquiring a toy, whereas invisible friends were often viewed as fulfilling a need for a relationship. Compared to children without imaginary companions, children with imaginary companions were more likely to be firstborn and only children.
A very special friend: the imaginary companion. Imaginary companions are common in childhood and adolescence. However, little research has been conducted on this phenomenon, and the significance and functions of imaginary companions remain poorly understood. Conceptions about the psychopathological consequences of having an imaginary companion predominate. This contribution provides an overview of the current slate of research and an analysis of five different functions a child's or adolescent's imaginary companion may fulfill. Developmental changes in reality-testing, fantasies, and creativity are linked with this phenomenon as are changes in socio-emotional development and in friendship conceptions. The positive, stimulating function of an imaginary companion for further development is emphasized.
The relationship between self-image and creativity was studied in primary school children. Earlier research points in two directions. Some researchers describe the creative child as well adjusted. Others provide a more nuanced picture in which less well-behaved children can also be creative. Three different measures of creativity were used in this study: the Unusual Uses Test, an activity questionnaire and a perceptual test (the Creative Functioning Test). A self-image inventory was used to measure participants' perceptions of their own skills, physical self, psychological health and relationships to others. The results showed no self-image differences between children with high and low creativity. The creativity measures were significantly related, with the exception of the flexibility dimension of CFT. One possible explanation is that CFT measured another aspect of creativity. This was illustrated in a cluster analysis in relation to self-image.