Do monsters dream? Young children's understanding of the fantasy/reality distinction

British Journal of Developmental Psychology (2004), 22, 293–310

2004 The British Psychological Society

Do monsters dream? Young children’s

understanding of the fantasy/reality distinction

Tanya Sharon

* and Jacqueline D. Woolley

Emory University, USA

The University of Texas, USA

Young children are often thought to confuse fantasy and reality. This study took a

second look at preschoolers’ fantasy/reality differentiation. We employed a new

measure of fantasy/reality differentiation—a property attribution task—in which

children were questioned regarding the properties of both real and fantastical entities.

We also modified the standard forced-choice categorization task (into real/fantastical)

to include a ‘not sure’ option, thus allowing children to express uncertainty. Finally, we

assessed the relation between individual levels of fantasy orientation and fantasy/reality

differentiation. Results suggest that children have a more developed appreciation of the

boundary between fantasy and reality than is often supposed.

The distinction betw e en fantasy and re ality is basic to human cognition, re flec ting a

fundame ntal ontological divide be tw e en the non-re al and the real. Children have

traditionally be e n tho ught to c o nfuse the boundary betw een fantasy and reality. Piaget

(1929, 1930) he ld that c hildren not o nly c onfuse fantasy and reality but the mental and

the physic al, dreams and re ality, and appearance and re ality. The influenc e of this

perspec tive is still felt in e arly childhoo d educ ation, media and c o mmon-sense be liefs of

adults (se e, e .g., Daw kins, 1995).

Ho w eve r, the view that children confuse fantasy and reality is at o dds w ith a large

body o f re se arc h sho w ing that c hildre n as young as three years are able to make various

o th e r

no n-re ality/ reality distinctions. Fo r ex ample , by thre e years of age, children c an

distinguish a mental e ntity, such as a thought o r an image , from the re al physical objec t

it re prese nts (Este s, We llman, & Woolley, 1989; Wellman & Este s, 1986). At about this

same age, childre n in their eve r yday talk disc uss the contrasts betw e e n to ys and reality,

picture s and re ality, and pre te nse and re ality (Wo olley & Wellman, 1990). They c an

trac k re al and pre te nd transformations c o nc urre ntly (Harris & Kavanaugh, 1993) and,

w he n their prete nd play is inte rrupte d, are able to flexibly ste p out of the pretense

mode then re turn to it (Go lumb & Kue rste n, 1996). In contrast to this rese arc h, the

www.bps.org.uk

* Correspondence should be addressed to Tanya Sharon, Department of Psychology, Mercer University, 1400 Coleman

Avenue, Macon GA 31207, USA (e-mail: sharon_t@mercer.edu).

293

smaller body o f w o rk specifically ex ploring childre n’s unde rstanding of the fantasy/

re ality distinc tion c onfirms Piaget’s view that children ofte n fail to o bserve the prop e r

bo undary. Fo r ex ample , Taylor and How e ll (1973) pre se nte d 3- to 5-year-old childre n

w ith both fantastical and natural pictures of animals, and asked children to state

w he ther the de picte d sc ene s could really happe n. Thre e -year-olds had c o nsiderable

difficulty differe ntiating re al from fantastical sce nes. Using a similar me thod, Samue ls

and Taylor (1994) fo und that c hildren w ere most c o nfuse d w hen eve nts w ere

pe rceive d as frighte ning. The se findings imply a role fo r emotion in young c hildre n’s

understanding of re ality status and suggest that their grasp of the fantas y/re ality

distinc tion may be so me w hat fragile. In another study (Moriso n & Gardne r, 1978), a

some w hat firmer g rasp of the distinction eme rge d in an o lder gro up of children

(kindergarte n thro ugh sixth grade), but pe rformance continue d to impro ve throug ho ut

the grade scho ol ye ars.

Emp irical researc h o n the fantasy/re ality distinc tion is thus co nsistent w ith Piaget’s

view that young c hildren have difficulty negotiating the bo undary betw ee n fantasy and

re ality. It is also co nsistent w ith surve y data show ing high levels o f belief in spe c ific

fantasy e ntitie s suc h as Santa Claus and the Easter Bunny (Clark, 1995; Pre ntice ,

Mano se vitz, & Hubbs, 1978; Rose ngre n, Kalish, Hickling, & Ge lman, 1994). It se ems

c le ar that children have stro ng leve ls of be lie f in e ntities that to adults are

unambiguously fantastical.

These various so urces thus pre se nt a somew hat inc onsiste nt picture of young

childre n’s understanding o f differe nt non-re ality/reality c ontrasts. To begin to make

se n se o f th e se findings, w e take a se c o nd lo o k at c hildre n’s fantasy/re a lity

differentiation. Our me thods w e re guided in part by the belief that the categorizatio n

tasks use d in pre vious re se arch may give an o ve rly simplistic reading of childre n’s

understanding, lending their respo nse s an appearanc e o f o ntological commitme nt no t

actually felt by them. We w ere also guided by the finding that c hildren often reveal

greate r c ate gory unde rstanding w he n the ir know ledge is asse sse d via inductio n tasks

that te st their ability to mak e appro priate inferenc e s (e .g. Ge lman & Markman, 1987). In

a typic al inductio n task, children are taught a no vel pro pe rty for an o bject and the n are

tested for the cate go rical range ove r w hich the y are w illing to gene ralize that prop erty.

This method has pro ven e x tre mely pro duc tive in re vealing the de pth and c omplex ity o f

childre n’s early c onc epts (e.g. Gelman, 1989; Kalish & Ge lman, 1992; Lopez, Gelman,

Guthe il, & Smith, 1992; Mandler & McDonough, 1996). Give n this, it is p ossible that

childre n w ho fail to label e ntities acco rding to adult cate gorie s of re al and pre te nd may

still rec o g nize differenc e s be tw ee n real and fantastical entitie s in te rms o f the ir abilities

and prope rties. This possibility has no t pre viously be en te ste d.

Prio r w o rk do e s ho w ever suggest that at least part of the ne ce ssary know ledge is in

plac e. Spe c ifically, rese arch has show n that yo ung c hildre n have clear ideas abo ut the

kinds of things real e ntities can and c annot do. By four o r five years o f age, they can

identify and differentiate re al entitie s o n the basis of the ir physical, p sycho logical and

biolo gical pro pe rties (We llman & Gelman, 1998). For ex amp le, childre n o f this age

kno w that living things occ upy spac e, have tho ughts and gro w . At the same time , many

fantasy figure s po sse ss c le ar no n-human abilities, suc h as the ability to travel great

distance s instantaneously. Children may thus ap preciate that the se prop ertie s are

unusual and—mo re importantly—non-human, before they ap prec iate that the e ntities

that po ssess them are nece ssa rily fantastical. In othe r w o rds, young children may

demo nstrate more ac c urate fantasy/re ality differentiatio n in the p rop e rties they

attribute to various entitie s than in the c ategories to w hich the y assign them.

294 Tanya Sharon and Jacqueline D. Woolley

The first goal o f the pre se nt study w as to ex plo re this possibility. Spe cifically, w e

inve stigate d the number and kinds of prop e r ties childre n and adults attribute d to a

range o f re al and fantastic al entitie s. Pro pe rties from four fo undational domains w ere

included (physica l, biological, so c ial and me ntal), be c ause previous re se arch has show n

that children’s kno w ledge de ve lo ps at differe nt rates in these domains, w ith physical

and soc ial kno w ledge de veloping be fore know ledge of biolo gy o r me ntal state s (e.g.

Care y, 1996; Hirshfield & Ge lman, 1994). Adult pro pe rty attributio ns w ere also assessed

in order to (1) c onfirm that adults disc riminate in the ir pro p erty attributions for real and

fantas tical entities, and (2) e stablish the adult patte r n of attributions acro ss do mains.

A se c o nd goal w as to obtain a mo re nuanc e d p icture of yo ung c hildre n’s

categorization o f re al and fantastic al e ntities. As in pre vious studies, w e had childre n

categorize a var ie ty o f e ntitie s. Ho w ever, inste ad of having c hildren simp ly cate gorize

entities as ‘re al’ o r ‘pre tend’, w e inc luded a third ‘no t sure ’ o ption to c apture

unce rtainties, w hich arguably c o uld be ve ry pre valent in this domain yet w o uld not be

capture d in the traditio nal so rt.

The third goal of this study w as to investigate the po tentia l role of individual

differe nc es in c hildr en’s fantasy/ re ality differe ntiation. Oddly, the ro le of individual

differe nc es in fantasy/ reality differentiation is rarely addressed (cf. Bourc hier & Davis,

2000; Johnso n & Harris, 1994; Wo olley, 1997). Some c hildren are much mo re incline d

than othe rs to engage in fantastica l p ursuits, such as p rete nding or having an imaginary

co mpanion (Taylor, 1999). Suc h a high fantasy o rientation (FO) could have great

po tential relevanc e to c hildren’s be liefs in fantastical figure s. On the one hand, a high

FO co uld inc rease be lief in fantastical entities. FO and a w illingness to be lieve in the

ex istenc e of e ntitie s for w hich the re is little empirical support have often bee n

co njo ined (Singer & Singer, 1990; Vyse, 1997). On the other hand, it is no t nec ess arily

the c ase that some o ne w ho enjoys and e ngages in fantasy frequently is also some one

w ho is confused abo ut w he re to draw the bo undary. In fact, just the o ppo site may be

true . A c hild w ho e ngages in fantasy a gre at de al may develop a mo re highly tune d se nse

of w hat is re al and w hat is no t pre cisely bec ause of his o r her play.

It is also p ossible that fantasy o rientatio n and understanding o f the fantasy/ r e ality

distinction are unrelate d. Data have be en sp arse and co ntradic tory, w ith so me

re se arche rs finding a p ositive relation (Singe r & Singe r, 1981) and others finding no

re lation (Dierker & Sanders, 1996–1997; Pre ntice

e t a l.

1978; Taylor, Cartw right, &

Carlso n, 1993). One rec e nt study aime d spe c ifically at this que stio n (Bouldin & Pratt,

2001) found that c hildren w ith an imaginary comp anion w e re mo re likely than c hildren

w ithout such a c ompanion to e ntertain the po ssibility that a briefly prese nte d image

w as a monster. Ho w ever, a substantial proportio n of c hildren w ithout imaginary

co mpanions also co nsidere d this po ss ibility, leading the autho rs to suggest that it is

individual differe nce s in c re dulity, rathe r than fantas y orientation per se, that leads to

fantas y/reality co nfusion. Rese arc h is c omplic ate d by the fact that, at pre sent, there

ex ists no single, validated sc ale o f fantasy o rientatio n. To e nable an initial e x plo ratio n o f

the issue in the c ur re nt study, a range o f fantasy orie ntation tasks w e re included.

To summarize, the go als of the curre nt study w ere (1) to dete rmine w he the r c hildren

differe ntiate be tw ee n re al and fantastic al entities in the pro p erties they attribute to

the m, (2) to pro be c hildren’s c ate gorization o f a range o f real and fantastic al e ntities,

and (3) to ex plore the influenc e o f gene ral fantasy orie ntatio n o n c hildren’s fantasy/

re ality differentiation. Although the re are indications in the literatur e that verbal and

behavioural me asure s can re ve al different leve ls of understanding o f the fantasy/ reality

distinction (e .g. Harris, Bro w n, Marriott, Whittall, & Harme r, 1991; Subbo tsky, 1997;

295Fantasy/reality distinction

Wo olley & Phelps, 1994), including both types of me asures w o uld have made o ur

alre ady lengthy study ex c essive ly pro lo nged. We the refo re cho se to use o nly verbal

me asure s, to fac ilitate comparison betw een o ur re sults and re sults o f tho se studies most

clo se ly re late d to ours (Mo r ison & Gardner, 1978; Samue ls & Taylo r, 1994; Taylor &

Ho w ell, 1973).

Method

Participants

Six ty-four pre sc hoo lers w e re te ste d, inc luding tw enty-tw o 3-ye ar-olds (6 male and 12

female,

= 40 months, range 33–47), nine tee n 4-year-olds (10 male and 9 female,

51 months, ran ge 48–58) and tw e nty-three 5-year-olds (13 male s and 10 females,

= 62

mo nths, range 60–68). Respo nse s from thre e p artic ipants w ere omitted, o ne due to

unc oo pe rative ne ss (a 4-year-o ld bo y) and tw o be cause of ‘ye s’ biase s (tw o 3-ye ar-o ld

bo ys). Most children attended a c hildcare cente r affiliated w ith a large unive rsity.

Informe d c onsent w as obtaine d fro m all pare nts. Adult participants w e re 38 adults

enro lled in an introduc tory psychology course and 27 additional adults atte nding the

same university.

Procedure

Children w e re teste d individually by o ne of three female ex pe rime nters in tw o separate

25-min sess ions. Se ssio ns w e re sp ace d 1–15 days apart (

= 5 days). Te sting involve d

three tasks and an inte rview assess ing the c hild’s level o f fantasy o rientatio n. The tasks

w ere a pro pertie s task, a c ategorizatio n task, and 3 min of fre e play w ith blo cks. In the

first session, children complete d half o f the prop e r ties task (3 entities), the blocks task

and the FO inte rview . In the se c ond se ssion, childre n co mple ted the se cond half of the

properties task (3 e ntities), the FO intervie w again and the cate go rization task. The FO

intervie w w as c omp lete d tw ice to provide a mo re ac curate asse ssment of c hildre n’s

fantasy orie ntation.

Properties task

Children w e re presented w ith c oloure d line draw ings of various e ntities. Pre vious

re search has suggeste d that childre n re spond to, and re aso n diffe r e ntly about, gene r ic

fantastical entities (such as mo nsters and ghosts) compared w ith sp ecific entitie s

associated w ith ce rtain e ve nts (such as Santa; Rose ngre n

e t a l.

, 1994; Sobe l & Lillard,

2001). Ac co rdingly, children w ere pre se nte d w ith six entitie s: 2 real entities, o ne

specific (Michae l Jordan) and o ne gene ric (a c hild of the same se x as the subjec t), 2

specific eve nt-related fantastical e ntities (Easte r Bunny and Santa Claus) , and 2 generic

fantastical e ntities (a monster and a fairy). Eac h picture w as pre sented individually, and

childre n’s rec ognitio n of the e ntity w as asc ertaine d. Recognition w as unive rsal w ith the

ex ceptio n of Michae l Jordan, for w ho m re c o gnition w as low e r (53, 50 and 90%for 3-, 4-

and 5-year-olds, respectively). Childre n w ho did not reco gnize Jordan w e re pro vided a

brief de sc ription (‘Michae l Jordan is a famous basketball player. He is very fast and

stro ng’).

For each e ntity, a se ries o f 12 yes/no que stions w e re asked re garding w he ther the

entity po sse sse d spe cific pro perties. There w ere three questions in e ach of four

296 Tanya Sharon and Jacqueline D. Woolley

domains: physical, biological, soc ial and me ntal (see Appendix ) Questions w e re

designe d to tap a bro ad range of pro pe rties, including both tho se that are nec e ssarily

true o f humans (e .g. biological p rope rtie s suc h as slee ping) and those that are no t

ne c essarily true but highly charac te ristic o f humans as childre n know the m (e .g. soc ial

propertie s suc h as e ating dinne r w ith one’s family). The o r de r o f questio ns w as fixe d.

The o r de r of e ntities w as partially co unterbalance d via e ight differe nt o rders of

prese ntation aro und the c o nstra int that o ne re al e ntity be pre sente d in eac h session.

Categorization task

Children w e re sh ow n thre e trays o f different colours. The ex pe rime nte r ex plained that

one tray w as for ‘r e al one s’, one w as for ‘pre tend one s’, and one w as for ‘one s you’re

no t sure about, o r you don’t kno w ’. The ‘not sure ’ tray w as alw ays in the centre;

plac ement of the o the r tw o trays w as c ounterbalanc ed. Nex t, children w e re show n

individual picture s of entities pre sente d in a random order w ith the que stio n ‘Who is

this?’ for sp e c ific figure s (e .g. Santa) o r ‘What is this?’ for gene ric ones (e.g. a mo nste r).

They w ere then asked, ‘Where doe s this o ne go ?’. Bec ause this task w as much less time -

intensive than the prop e rtie s task, w e w ere able to add mo re entities: tw o re latively

familiar figures (a clow n and a magician), tw o less familiar figure s (a knight and Robin

Ho o d) and a c o mic -bo o k figure (Supe r man). Also inc lude d w e re tw o ro ughly

co mparable no n-human e ntities, one fantastical (dragon) and one re al (dinosaur).

Rec ognition o f the additional e ntities w as generally ve ry high.

Fantasy orientation assessment

The Fantasy Orie ntatio n asse ssme nt inc luded fo ur se parate measures of fantasy

orie ntation use d in pre vio us re search: a thr e e -p ar t interview , and 3 min o f free play

w ith blo c ks.

It included Taylor and Carlson’s (1997) Impe rso nation Interview , a

shorte ned versio n o f their Imaginary Companion (IC) Interview and a subse t o f

questio ns draw n from Singer’s IPP (Imaginative Play Pre dispo sition) sc ale (Singe r , 1973;

Singer & Singe r, 1981). In a large study utilizing 20 differ e nt pote ntial measures o f

fantas y orientation (Taylor & Carlso n, 1997), Singe r’s IPP loaded significantly and

he avily o n a ‘fantasy engagemnnent’ fac tor. Measure s o f the c hild’s po ss ess ion of an

imaginary c o mpanion and the ir impe rso nation be haviour also loade d highly and

significantly.

The testing situation at the pre sc hoo l pr e c lude d r e c ruiting pare ntal input o n the FO

measure s, w hich can be a he lpful w ay of c ross-validating c hildren’s r e spo nse s. In lieu o f

this, w e administe r e d e ac h o f the intervie w FO me asure s tw ice , once in e ach se ssio n;

sc ore s from bo th se ssio ns w e re the n summe d. In this w ay, c hildren w ho w ere mo re

co nsistent in their fantasy orie ntation (w ho gave fantastical response s both time s)

ac hieve d a highe r sc ore on the individual measure than childre n w ho w e re less

The exception was Robin Hood, who was not recognized by a sizea ble minority of children. Robin Hood was accordingly

dropped from analysis. The magician was also dropped du e t o ambiguity regarding his status even among adults: Many adults

would say that a magician is a real person while holding that magic itself does not ex ist. The only other entity that occasioned

any difficulty in recognition was the knight, which was sometimes identified with more general military figures (soldier, guard,

army). All other entities were recognized by all but 1 or 2 children. In the few cases where children did not spontaneously

identify the entity, or identified it incorrectly, a short description was provided. Thus, children who labelled the picture of a

knight a ‘soldier’ were told ‘This knight fights his enemies with his sword.’

Children’s scores on the blocks task were not significantly correlated with either Singer’s IPP or Taylor and Carlson’s

Impersonation interview, nor were scores related to any of the dependentmeasures. The blocks task was accordingly dropped

from the analysis and will not be discussed further.

297Fantasy/reality distinction

co nsistent (giving so me fantastical and so me no n-fantastical response s). Children w ho

gave c o nsiste ntly no n-fantastical response s sc ore d the low est.

The Imp e rso natio n Inte rview (Taylor & Carlso n, 1997) co nsisted of three questions

asking w hethe r children e ver prete nded to be (a) an animal, (b) a different person, and

answ er for w hich they c ould also give at least o ne ex ample; thus, score s after bo th

sessions c o uld and did range fro m 0 to 6 (

= 3.15).

In the Imaginary Co mpanion Inte rview , the term ‘pre tend friend’ w as first e x plained,

and c hildren w ere aske d w he ther they had a ‘pre te nd friend’. For childre n w ho

answ ered ye s, a series o f nine follow -up questions w e re asked co nce rning issue s such

as the friend’s name and gende r; w he re the IC lives and sle e ps; and w he ther the IC w as

a to y o r to tally pre te nd. The follow -up que stio ns allow ed us to e valuate the

substantiveness o f po sitive re sp o nse s. ‘Yes’ re sponse s that clearly re f e rred to a sp ec ific

re al c hild (e.g. a co usin) o r pe t, o r for w hich no details c ould be provide d, w ere sc ore d

as 0. Ne gative respo nse s at both se ssions w ere sc ore d 0. A single substantiated positive

re spo nse w as co de d 1; substantiate d, po sitive re sponse s at both se ssions w e r e sc ore d 2.

The modal response on this me asure w as 0; almo st 60% o f c hildren did not re port an

imaginary co mpanion, w hile 40% did. This pro p ortion is comparable w ith pre vious

re search (Taylor & Carlson, 1997).

Four que stio ns from Sin ger’s (1973; Singer & Singer, 1981) IPP scale w e re use d,

re garding children’s favourite game and favourite toy, w hethe r they like to talk to

the mse lves in bed at night and w hat they think abo ut be fore they go to sleep .

Re sp o nse s w e re coded, fo llow ing Singe r’s (1973) guidelines, as either fantasy-orie nte d

(e.g. monste r game), w hich re c e ived a score o f 1, or re ality-orie nte d (e .g. checkers),

w hich re c e ived a score of 0. Re spo nse s that suggeste d an imaginative c o ntent w ithout a

cle ar fantasy e leme nt (e .g., playing fireman) w ere co ded 0.5. Scores on this measure

afte r both se ssions co uld ran ge from 0 to 8. Actual obtained sc o res range d from 0 to 6

(

= 2.68).

Reliability of FO measures

On e e xperime nte r sc o red all pro toco ls. A differe nt ex pe rimenter re-sc ored 45% o f the

pro to c o ls. The inte r-rate r agre e ment w as 93–98% fo r the 3 que stio ns in the

Imp e rso nation Intervie w ; 93% fo r the Imaginary Comp anio n interview ; and 78 to

89% for the 4 questions in Singer’s IPP. Individual c hildre n’s respo nses to the FO

intervie w me asure s ac ross the tw o se ssions w ere also assesse d and found to be largely

co nsistent. Agree ment for the 3 questions in the Impersonation Inte rview ranged from

73 to 79%; for the Imaginary Co mp anion interview , the figure w as 82%; and fo r the 4

questions in Singer’s IPP, agree me nt ranged from 74 to 85% e x c ep t for the ‘sle ep’

question, for w hich agree ment w as 63%.

Testing of adults

The pro perties task w as administe red to 38 adults as a simple four-page que stionnaire

that to o k ap prox imately 10 min to c omplete . Although w e felt c o nfident that adults

w ould agree w ith o ur intuitions about the reality status of the entities (e .g. that Michael

Jo rdan is real and that Santa Claus is no t), w e confirme d o ur intuitions o n a separate

sample o f 27 adults. The se adults w e re given a list of the entities and asked to c hec k

w he ther the y w e re real o r no t (o r w he the r they w e re unsure). For the 11 e ntitie s used

in the final analysis, the judgments of the se adults confirme d o ur intuitions 97% of the

time (288 out o f 297 judgments).

298 Tanya Sharon and Jacqueline D. Woolley

Results

FO measures

Bec ause o ne o f the me asures of fantasy orientation w as catego rical rathe r than

co ntinuo us, the relations betw ee n the me asure s w ere asse sse d by a non-parametric

measure o f co rre lation, Spe arman’s rank c orrelatio n c o efficient. Results sho w ed a

significant relation betw een Taylor and Car lson’s Im p erso nation Inte rview and both

Singer’s IPP Interview ( = .304,

< .05) and the IC Intervie w ( = .324,

< .05). The re

w as also a marginally significant relationship betw e en the latte r tw o me asures, = .217,

= .09. Acco rdingly, the se thre e measure s w ere c o mbine d to cre ate an ove r all FO

sc ore . To c o rre ct fo r the different sc ales o n the dif fere nt me asures, sc ore s w ere first

co nve rted to

score s and then summed. The resulting sc ale had a me an of 0 .28 and

ranged be tw ee n 4.09 and 3.98. Inspec tion o f the distribution of sc ore s show ed a

break just below the median sc o re. Thus, c hildren w ho sc o red be low the median w e re

classified as low FO (

= 29) and children at or abo ve it as high FO (

= 32). The

distribution o f high and lo w FO did not differ significantly acro ss the age gro ups,

(2)

= 1.97,

= .37, nor by gender

(1) = .85,

= .39.

Categorization of the entities

Pre liminary analyses sho w ed no main e ffects or interac tions involving orde r o r ge nde r;

the se variable s w ere dro ppe d from subse quent analyse s. The first se t o f analyse s

inve stigate d childre n’s fantasy/re ality differentiation as me asured in their c ategorization

of the vario us entities. Consistent w ith pre vious rese arch, childre n performe d po orly.

Few e r than half (46%) of the ir total cate gory judgments w ere co rre c t. A 3 (age) 2

(FO) 2 (entity typ e) mix ed analysis o f varianc e w as c onducte d on the p ropo rtio n of

co rrec tly cate go r ize d e ntities, w ith age and FO as be tw e en-subjec ts variables and e ntity

typ e as the w ithin-subjects variable.

Results sho w ed a main e ffec t of FO (1, 51) = 6.38,

< .05. Children w ith a high FO w e re more ac curate in the ir c ate gory judgments than

w ere low FO c hildre n (54 vs. 40%).

There w ere also main e ffec ts of age,

(2, 51) = 8.74,

< .001, and e ntity type,

(2,

51) = 10.54,

< .001, qualified by a significant interac tion be tw e en them,

(2, 51) =

10.54,

< .0005. Inspe ction of the me ans sho w ed that correct cate go rizatio n of the re al

entities inc rease d w ith age (

= 33, 60 and 74% for 3-, 4- and 5-year-o lds, respec tive ly),

w hile c orre ct c ate gorization of the fantastic al e ntitie s sh ow e d no improveme nt (

= 34,

39 and 39%; See Fig. 1). Thus, all age gro ups p erfo rme d at c hance levels in c ate gorizing

the fantastic al entitie s, and 3-ye ar-o lds w e re e qually po or at catego rizing the re al

entities. The po or sc o res o f 3-year-olds canno t be attribute d solely to the ir lack of

familiarity w ith Jordan, as the same p attern w as o btained w ith the Michae l Jo rdan item

ex clude d. Ho w ever, inspe ction of childre n’s patte rns o f re sponse to individual e ntitie s

(disc usse d sho rtly) show s that, rathe r than be ing rando m, the ir p o o r o ve rall

performanc e re flec ted the fact that so me fantastic al entities w e re syste matically judged

as fantastic al, w he reas o thers w e re syste matically judged as real.

Children’s use of ‘not sure’

Before ex amining these data, w e re moved the entities that childre n had difficulty

Proportion correct was used in place of raw scores to correct for the baseline difference in number of real versus fantastical

entities. All analyses in this paper conducted on proportional scores were also ru n with an arc-sine transformation on the

dependent variables. The pattern of results was identical in every case.

299Fantasy/reality distinction

re co gnizing (Michae l Jordan and the knight). This w as do ne to ensure that childre n’s

use of the ‘not sure’ option reflecte d unc ertainty regarding the se e ntities’ fantasy/ reality

status rather than the ir identity. We found that children used the ‘no t sure’ o ptio n

ex tensively. Virtually all e ntitie s w ere assigned to this category by at least some children

at eac h age , and use freque ntly range d ove r 30%. Overall, ac ro ss the age gro ups and

entities, 21% o f assignme nts w ere to the ‘not sure ’ box . The se findings thus reveal an

impo rtant element o f unc e r tainty in children’s fantas y/ reality judgme nts.

Category judgments for individual entities

Among 3-year-olds, the only e ntity for w ho m children’s patte rn o f re spo nse s differe d

significantly from a c hanc e level o f 33%, as me asure d by Pearso n c hi-square te sts, w as

the knight (Table 1). Further, the ir c onse nsus o pinion w as inc orre ct. Four-year-olds’

cate go ry cho ices w ere more c onsiste nt. Among the real entities, response s fo r the

child, c low n and Michae l Jo rdan w ere all syste matically differe nt fro m c hanc e (

< .01),

w ith the majority co r re c tly assigning re al status. Thus, 4-year-olds de monstrate d a clear

understanding of the re ality status o f both spe cific and gene ric human figure s. Among

fantasy figures, categorization choice s fo r Supe rman and Santa w ere significantly

different from chanc e (

< .05), w ith the majority o f children correctly c ate go rizing

Supe rman as prete nd but incorrec tly catego rizing Santa as re al.

By five ye ars of age , children made syste matic, if some times erro neo us, c ate gory

judgments for all but tw o entities (fairy and dragon). The majority o f this age group

co rrec tly cate go rized all of the re al entities as real and the monster and Supe rman as

pre tend (

< .05). Ho w eve r, the majority also incorrectly c ategorize d the tw o eve nt-

re late d fantastical e ntities—the Easte r Bunny and Santa—as re al (

< .001.)

Property attributions

The se cond set of analyses conc erned the attribution of various pro pe rtie s to the

different e ntities. The most basic que stion w as w he the r childre n w o uld show any

discrimination in their p rope rty attributions base d on their

o w n

notions o f w hat w as

Figure 1.

Correct categorization of entities by age and entity type.

300 Tanya Sharon and Jacqueline D. Woolley

Table 1.

Children’s categorization choices for each entity (%), by age

Real Fantasy

Child Michael Jordan Clown Knight Dinosaur Eas ter Bunny Santa Fairy Monster Superman D ragon

3-year-olds

Real 50 28 39 17* 33 44 44 50 39 33 27

Not sure 22 33 28 22 28 28 33 33 28 17 20

Pretend 28 39 33 61 39 28 22 17 33 50 53

4-year-olds

Real 71** 94*** 71** 35 29 35 76*** 41 29 35* 18

Not sure 24 6 24 35 41 35 12 29 24 6 24

Pretend 6 0 6 29 30 29 12 29 47 59 59

5-year-olds

Real 82*** 73*** 86*** 59* 64** 81*** 71*** 52 24* 14** 27

Not sure 9 14 0 18 23 10 14 14 14 18 18

Pretend 9 14 14 23 14 10 14 33 62 68 55

Patterns of responding that are significantly different from chance by a chi-square test are indicated: < .10; * < .05, ** < .01; *** < .001.

301Fantasy/reality distinction

re al or fantas tical. To pro vide the c leare st pic ture po ssible, entities categorize d as

‘unc ertain’ w e re not inc luded in this analysis. Average attribution sc ore s in e ac h o f the

four domains w e r e c alculated for e ac h child based on his or he r o w n category

judgments of ‘re al’ and ‘fantastic al’. Eight se p arate sc o res resulted, e ac h o f w hich c o uld

range from 0 to 3 (e .g. ave rage attributio n o f so c ial prop erties for e ntities judge d re al,

average attribution o f biological pro perties fo r e ntities judged fantastic al, etc .). As

childre n assigne d different numbers of entitie s to eac h c ategory, score s w e re c alculate d

as ave rage pro po rtions rathe r than ave rage raw sc o res.

Sc o res w ere e ntere d into a 3 (age ) 2 (FO) 2 (e ntity classificatio n) 4 (pro pe rty

typ e ) mix e d ANOVA. Re sults sho w ed only a main e ffect o f attributions,

(1, 45) = 8. 9,

< .01. Children grante d mo re human-lik e pro p erties of every type to e ntitie s the y

classified as real than to those the y classified as fantastical; ho w ever, post-hoc Sc he ff e´

tests sh ow e d that the o nly significant difference w as in the soc ial do main (71%vs. 50%).

This suggests that children’s infere nc es abo ut the so c ial p rope rtie s of an e ntity may play

a leading role in the ir judgme nts about the e ntity’s re ality status.

The ne x t ste p w as to e xamine c hildren’s pro pe rty attributions to e ntities as define d

by the adult catego rie s of re al and fantastical. The adult data w e re analysed first to

establish the adult patte rn o f attribution, follow ed by analyses of the children’s data that

also took into acco unt their age and fantasy orientation.

A repe ated-measure s ANOVA o n adult’s attributions by pro perty and entity type

produc ed a significant main effec t of attributions,

(37, 259) = 124.08,

< .0001. Po st-

hoc S c heffe´ tests sho w ed that adults endorse d significantly more pro pertie s o f e ve ry

typ e for re al e ntities c o mpared w ith fantastical entities (all

< .0001) (se e Fig. 2A). In

addition, adults did no t discriminate in the types of human-like propertie s the y

attribute d to re al entities; the y w ere all e ndo rse d at equally high le vels. Ho w ever, adults

did differentiate in the ir pattern of attribution for fantastical entities: Human-like so c ial

properties w ere attributed significantly le ss o ften to fantastical entities than any othe r

typ e o f pro pe rty (

< .05). Attribution of the othe r type s of properties to fantastical

entities w as co mparable .

To assess children’s patterns o f attribution by pro pe rty and e ntity ty p e,

a 3 (Age)

2 (FO) 2 (e ntity type ) 4 (pro pe rty typ e ) mix ed ANOVA w as conduc ted, revealing

main effec ts of age ,

(2, 55) = 5.95,

< .005, and FO,

(1, 55) = 10.56,

< .005. Thr e e-

year-o lds attributed few e r pro pe rtie s o n average than did 4- or 5-ye ar-o lds (

= 1.7, 2.0,

2.1 out of 3, resp ective ly,

< .005), and high FO children attributed more pro perties (

= 2.0) than did low FO children (

= 1.8). There w as also a significant main effec t o f

attributions,

(7, 385) = 46.78,

< .0001, qualified by a significant age -by-attribution

interaction,

(14,385) = 4.32,

< .0001.

To de termine the cause of the age by attribution inte rac tion, post-hoc Scheffe´ te sts

w ere c onducted o n the pattern of attributio ns sep arately fo r e ach age. The results (see

Fig. 2B–D) re ve aled that the 5-year-olds c lose ly matc he d the adult patte rn of

attributions: Among 5-year-olds, real entities w e re attribute d more human-like p rope r-

ties of each type than w ere fantastical e ntities,

< .0001, w hile attributions o f the

differe nt types of pro pe rties w ithin e ach entity type did not differ. Four-year-olds

show ed a similar patte rn, altho ugh the ir differentiation betw ee n re al and fantastic al

entities w as not as co mplete : The y ac co r de d significantly more human-like so c ial (

2.6) and p hysic al propertie s (

= 2.4) to real entities c ompare d w ith fantastical e ntities

Because a large proportion of children did not spontaneously recognize Michael Jordan, we conducted an initial analysis

to determine whether the child’s recognition had any effect on patterns of attribution. The result was negative F(1, 56) = .04,

p = .83.

302 Tanya Sharon and Jacqueline D. Woolley

Figure 2.

Average number of human-like properties endorsed by type of entity, property type an d age

(maximum = 3). indicate endorsements for real entities, indicate endorsements for

fantastical entit ies.

303Fantasy/reality distinction

Figure 2.

Continued

304 Tanya Sharon and Jacqueline D. Woolley

(

= 1.5 and 1.3,

< .005) but did not discriminate significantly be tw ee n re al and

fantas tical e ntities in the ir attributions of me ntal pro pe rtie s. They sho w ed a no n-

significant tre nd to w ards differentiating in the bio logical do main (

= 2.4 and 1.7),

.08. In co ntrast, 3-ye ar-o lds show ed no systematic differe ntiation by pro pe rty type,

attributing real and fantastical e ntitie s roughly equivalent numbe rs of pro pe rtie s o f e ach

typ e .

Finally, there w as also a significant thre e-w ay interaction be tw ee n age, FO and

attributions,

(14, 343) = 1.81,

< .05. Inspe c ting c hildren’s me an attributions

suggeste d tw o source s o f this interaction. First, high-FO 3-year-o lds w ere relatively

enthusiastic in the ir endorse me nt of social propertie s for fantastical e ntities c o mpared

w ith the othe r groups, (

’s = 2.4 vs. 1.4–1.7). Seco nd, 4-year-o lds’ attributions o f

human-like pro pe rties to re al e ntitie s re ve ale d an intere sting gap based on fantasy

orie ntation: High-FO 4-ye ar-o lds attribute d p rope rties similarly to 5-year-olds o f both

orie ntations o n 3 of the 4 pro perty typ e s (biolo gical, so c ial and me ntal), w he reas low -

FO 4 year-olds show ed a much low e r leve l of attributions and re se mbled the 3-year-olds.

This finding sug ge sts a po ssible advantag e of a high FO in making the fantasy/ reality

distinction.

Discussion

The mo st striking finding o f this rese arc h w as the c o ntra st be tw ee n c hildren’s relatively

inacc urate c ategory judgments and the ir relatively adult-like pro p erty attributions. Both

4- to 5-year-old childre n and adults differentiate d clearly be tw ee n real and fantastical

entities in the ir pro pe rty attributions, although childre n corre ctly categorized the

fantas tical entities only one third of the time. Furthe r, in the ir cate g ory judgments,

childre n often assigne d the ‘unc ertain’ cate go ry. This suggests that, rathe r than be ing

ac tively misguide d, young c hildre n are so metime s simply unsure about an entity’s

re ality status. Finally, an intriguing re lation w as found be tw e e n a high fantasy

orie ntation and more adult-like catego ry judgments for bo th re al and fantastic al entitie s.

Category judgments

Consistent w ith previo us re se arc h (Samuels & Taylor, 1994; Taylor & Ho w ell, 1973) ,

young c hildren in o ur study freque ntly failed to categorize entitie s corre ctly in te rms o f

the ir fantasy/ reality status. Ove rall, the y w e re co rre ct less than half the time . Thus, in

terms of assign ing e ntities to the correc t category, childre n of all age s in this study

performed po orly. How e ver, tw o additional analyses point to a mo re comp lex picture.

Children’s use of ‘not sure’

Pre vio us rese arch has c ommo nly offered c hildre n o nly tw o respo nse o ptions, ‘real’ and

‘pre tend’. If c hildren are unc e rtain about the status of a fantasy f igure , this w ill pro duc e

an e rro ne ous pic ture of c hildren’s beliefs, e ither inflating or de flating e stimate s of be lief

leve ls. Our proc edure allow e d us to capture children’s unc e rtainty regarding these

entities. When ‘no t sure’ re spo nse s are taken into ac count, w e find that fo r only a few

of the entities did the majority of c hildren at a given age ac tively assign the w ro ng

status. Thus, w he n given the o ppo rtunity, many c hildre n ackno w ledged the ir

unce rtainty regarding the re ality status of many entities. This is a very differe nt kind

of confusion than co nfidently holding a belief in the inc orrec t re ality status. A c hild w ho

305Fantasy/reality distinction

ex press es he sitatio n w hen asked w he the r a mo nste r is re al o r prete nd has not yet

achieved an adult understanding o f the insubstantiality of monste rs, but ne ithe r is he or

she c ommitted to a be lief in the ir re ality.

Category judgments for individual entities

The e ntities that pro duc e d the most c onsiste nt miscategorization (e.g. Santa Claus

among 4-year-olds, and S anta and the Ea s ter Bunny amo ng 5-ye ar-olds; see Table 1) w e re

largely those for w hich p are ntal and so c ial suppo rt is most strong (Prentice

e t a l.

, 1978;

Ro se ngre n

e t a l.

, 1994). The role o f so cial factors w as furthe r re ve ale d in the differenc es

in children’s be liefs ac ross the age groups. Of the 11 entities included in the

cate go rization task, 8 sh ow e d clear developme ntal tre nds, and 6 of the se w e re in the

‘c orre ct’ direc tion (i.e. inc reasing cate gorizatio n as real for the child, c low n, knight and

dinosaur, and inc re asing c ate gorization as pre te nd for the monste r and Supe rman).

On ly tw o entities sho w ed a clear tre nd in the ‘inc orre c t’ direction, and these tw o (the

Easte r Bunny and Santa Claus) are the fantastical e ntities that enjoy the mo st parental

and so c ial suppo rt (Rose ng ren

e t a l.

1994).

This patte rn o f re sponse s unde rsc ore s the importanc e o f the larger so cial co nte x t in

the developme nt o f c hildre n’s be liefs. The view of c hildren as highly cre dulous has

o ften failed to ac kno w ledge the e x tent to w hich c hildren are ex pec te d and enc ourage d

to hold c e rtain fantastical beliefs (c f. Woolle y, 1997). It sh ould be noted that Piaget,

w ho se w o rk c o ntribute d significantly to this view , made a stro ng distinctio n be tw ee n

childre n’s individual magical beliefs and the magical be liefs o f soc iety as a w hole

(Piaget, 1930). Yet these se e mingly paradoxic al trends—of both inc reasing ac c uracy

and incre asing erro r—c an be trac ed to children’s de ve loping kno w ledge o f and

ex pe rienc e in the w orld. Childre n learn in the co urse of a sc hoo l day that dino saurs

lived aeons ago, but the y also w rite letters to Santa from the se same classro oms. Pare nts

and o ther adults impart factual know le dge, but they also ac tive ly foster partic ular

fantastical be liefs (Clark, 1995). Thus, both the deve lopment of beliefs considered

c orre c t (e .g. dino saurs are re al), and the simultane ous de velo pme nt o f be liefs

co nsidered incorrec t, but age-appro priate (e.g. Santa is re al), o riginate in w hat children

are taught. Childre n may be more w illing to e ntertain fantastic al beliefs than are most

adults, but the y are also mo re uniformly e ncouraged to do so . The se data thus sugge st

that it is inc orre ct to view c hildren as generically c re dulous (Daw kins, 1995). The y did

not ye t make the same fantasy/ reality assignme nts as adults, but ne ithe r did they fail to

differentiate at all. Rather, their most common miscategorizations c oncerned spec ific

fantastical e ntities that yo ung children are gene rally enc ouraged to be lieve are real.

Property attributions

In c ontrast to their c ate gory judgments, children’s patterns o f prop e rty attribution w e re

similar to those o f adults, and reveale d an intriguing deve lop mental pro gression. Five-

year-o lds w ere identical to adults in their p attern of attributions ac ross the different

typ e s of prope rties (p hysic al, soc ial, e tc.). Four-ye ar-olds differe ntiate d be tw een

fantastical and re al e ntities in their attributio n o f social and physic al pro pe rtie s, w hile

failing to dif ferentiate significantly betw e e n the tw o in their attribution o f biolo gical and

me ntal pro pe rties. Whe n pro pe rty attributions w ere analysed in te rms o f the individual

child’s o w n c ategories o f real and pretend, 3-ye ar-o lds pe rforme d like the o lder

childre n, c o nsiste ntly attributing more human-like so cial pro perties to entities they

classified as real co mpared w ith e ntitie s they c lassified as pre te nd.

306 Tanya Sharon and Jacqueline D. Woolley

The develo pme ntal change in c hildren’s patterns o f prope rty attributions suggests an

important ro le for know ledge acquisition in the developme nt o f the fantasy/ reality

distinction. Children disc riminated first betw e e n the pro pertie s o f re al and fantastical

entities in the tw o do mains in w hich the y are apt to have the most kno w ledge (i.e. the

physical and the social; see, e .g., Fivush, 1997; Spe lke, Guteil, & Van de Walle, 1995).

The so c ial domain se eme d e sp ecially important at all ages. Adults grante d soc ial

propertie s to fantastical entities least o ften. The first sign o f differentiation amo ng the 3-

year-o lds in our study w as in the gre ater number of so c ial prope rties the y attributed to

entities the y c o nsidere d real c o mpared w ith tho se they considere d fantastical. Only

later did c hildre n discriminate in the domains in w hich they are still re lative no vice s

(i.e. the do mains o f bio logy and mind; se e , e .g., We llman & Ge lman, 1998). This re sult

undersc ore s the far-re ac hing effec ts of this fo undational kno w ledge.

Eve n w ith suc h kno w ledge , childre n face a c onside rable challenge, as they must still

re c ognize w hat implications the p rope r ty informatio n has for re ality status. Fantastical

entities may be partic ularly difficult be cause the y have c ontradictory pro pe rties. They

ofte n possess many attributes that are basic (e.g. facial featur e s) or highly typical (e .g.

Santa w ears c lothes and has a w ife) of re al e ntities. At the same time, the y po ssess

attribute s that are no t o nly unusual but also imp ossible for real entitie s (e .g. S anta also

flies in a sleigh). In a study on young c hildren’s unde rstanding of gende r, Ge lman,

Collman, and Macc oby (1986) found that c hildren w e re notably bette r at c orre c tly

inferring pro p erties, given a cate gory label, than the reve rse. Inferring the c o rre c t

category w hen given contra dictory pro pe rty information w as e spe c ially difficult. Yet

this is pre c ise ly the challenge that fantastic al entitie s raise . Childre n must come to

re c ognize, for examp le , that an entity may po sse ss many human-typical so cial

propertie s, but even one no n-human biological pro pe rty is e no ugh to c onsign it to

the re alm of fantasy.

The co ntrast betw ee n c hildren’s cate gory judgments and pro perty attributions has

another develo pme ntal implication. It suggests that man y pre sc ho o l-age childre n may

have an implicit c ategory of entities the y do not ye t labe l p rete nd, ye t to w hich the y

grant few human-like pro pe rties. This mix ed category co uld c onc e ivably se rve as a

transitional c ategory as children develop the adult fantas y category. As children learn

more abo ut the propertie s that are nec e ssary and typical of humans, the co ntrast

betw een the pro pertie s of the supposedly real e ntitie s and tho se o f the truly re al w ill

incre ase; this c o uld he lp c hildren re co gnize that so me of the e ntities are no t real but

imaginary.

Effects of fantasy orientation

Our final finding w as that a high FO had a po sitive effec t o n c hildre n’s c ate g orization of

both re al and fantastical e ntities: At all age s, high-FO children w e re more ac curate than

w ere the ir low -FO pe ers. This is an e x c iting finding, as there is ve ry little information in

the literature o n the ro le that individual differenc es play in childre n’s understanding of

the fantasy/ reality distinctio n. One possible e x planatio n is that high-FO c hildren spend

more time e ngaged w ith the fantasy w orld and the reby ac crue more kno w ledge about

it, se eing first hand bo th the limits and the p o ssibilities o f imagination. It is also possible

that these child ren’s intere st and e ngage ment in fantasy e licits mo re instruction from

parents, for ex ample discussion of the differe nc es betw e e n real friends and imaginary

one s. More w ork is so rely nee ded on this issue .

307Fantasy/reality distinction

Conclusion

Our re sults su ggest the follo w ing te ntative ske tc h o f th e p o ssible c ourse o f

deve lopme nt o f the fantasy/re ality distinctio n: The ve ry yo ung c hild may initially be

some w hat unsure abo ut attributing human -like prope rties to various e ntitie s. With

ex pe rienc e , childre n acquire inc re asing kno w ledge abo ut everything in their w orld—

bo th about real entities and the ir pro pertie s, and about suc h so cially sup porte d myths

as Santa Claus and the Easter Bunny. Thus, the re is the simultane o us developme nt o f

beliefs c o nsidered c orre ct (e.g. dinosaurs are real) and of beliefs c onside red inc o rre c t

but age-ap pro p riate (e .g. Santa is re al). But at the same time, as c hildre n be lieve in the

re ality o f fantasy figure s, o r are unable to say w ith c e rtainty that the y are prete nd, they

tre at the m very differently fro m re al e ntities in te rms o f the prope rties and abilities they

are w illin g to grant. In this w ay, c hildren se e m to plac e fantastic al e ntities in a separate

cate go ry—ne ithe r unquestionably re al nor pretend, but so mew here in be tw e e n. This

cate go ry co uld the n form a natural bridge to the adult cate go ry of fantastical entities.

Thus, rathe r than having misplace d the bo undary betw e en re al and fantastical entities,

young c hildren are still in the pro cess o f active ly c o nstruc ting it.

References

Bouldin, P., & Pratt, C. (2001). The ability of children w ith imaginary c ompanions to differentiate

betw een fantasy and reality. Br itish Jo u r n a l o f De v e lo p m e n ta l Psy c h o lo g y , 19, 99–114.

Bourchie r, A., & Davis, A. ( 2 000). Individual and developme ntal differenc es in children’s

understanding of the fantasy-reality distinction. British Jo u rn a l o f De v e lo p m e n ta l Ps y c h o lo g y ,

18, 353–368.

Boye r, P. ( 1997). Further distinctions betw ee n magic, reality, and re ligion. Ch ild De v e lo p m e n t,

68, 1012–1014.

Carey, S. (1996). Cognitive do mains as mo des o f thought. In D. R. Olson & N. To rrance (Eds.),

Mo d e s o f th o u g h t: Exp lo r a tio n s in c u ltu re a n d c o gn itio n (pp. 187–215). New York:

Cambridge University Pre ss.

Clark, C. (1995). Flig h ts o f fa n c y, le a p s o f fa ith : Ch ild re n ’s m y th s in c o n te m p o ra r y Am e ric a .

Chicago : University of Chicago Press.

Daw kins, R. (1995). Putting aw ay c hildish things. Ske p tic a l In qu ire r , January/Fe bruary.

Dierker, L. C., & Sanders, B. (1996–1997). Develop mental differenc e s and individual differe nc es

in children’s ability to distinguish reality from fant asy. Im a g in a tio n , Co g n itio n , a n d

Pe rs o n a lity, 16, 25–49.

Estes, D., Wellman, H. M., & Woo lley, J. (1989). Children’s understanding o f me ntal phe nome na.

In H. Re e se (Ed.), Ad v a n c e s in c h ild d e v e lo p m e n t a n d b e h a v io r (pp . 41–86). Ne w York:

Ac ademic Press.

Fivush, R. (1997). Eve nt memo ry in e arly childhoo d. In N. Co w an (Ed.), Th e d e v e lo p m e n t o f

m e m o ry in c hild h o o d ( pp. 139–161). Ho ve , UK: Psycho lo gy Press.

Gelman, S. A. (1989). Children’s use of c ategories to guide biological inference s. Hu m a n

De v e lo p m e n t, 32, 65–71.

Gelman, S., Co llman, P., & Mac coby, E. (1986). Inferring prop erties from c ategories ve rsus

inferring categories from prop erties: The c ase of gender. Ch ild De v e lo p m e n t, 57, 396–404.

Gelman, S . A., & Markman, E. M. (1987). Yo ung children’s induc tions from natural kinds: The role

of c ategories and appe arance s. Ch ild De v e lo p m e n t, 58, 1532–1541.

Golumb, C., & Kuersten, R. ( 1996). On the transition from prete nse play to re ality: What are the

rules o f the game? Br itish Jo u rn a l o f De v e lo p m e n ta l Ps y c h o lo g y , 14, 203–217.

Harris, P. L., Brow n, E., Marriott, C., Whittall, S., & Harme r, S. (1991). Monsters, gho sts, and

308 Tanya Sharon and Jacqueline D. Woolley

w itches: testing the limits of the fantasy–reality distinc tion in young c hildren. British Jo u rn a l

o f De v e lo p m e n ta l Ps y c h o lo g y , 9, 105–123.

Harris, P. L., & Kavanaugh, R. D. (1993). Yo ung ch ildren’s understanding o f p rete nse .

Mo n o g r a p h s o f th e So c ie ty fo r Re sa r c h in Ch ild De v e lo p m e n t, 58(1, Serial No. 231).

Hirschfe ld, L. A., & Ge lman, S. A. (Eds.). (1994). Ma p p in g th e m in d : Do m a in s p e c ific ity in

c o g n itio n a n d c u ltu re . New York: Cambridge University Press.

Johnson, C., & Harris, P. L. (1994). Magic: Spec ial but no t e xc luded. Br itish Jo u r n a l o f

De v e lo p m e n tal Ps y c h o lo g y , 12, 35–51.

Kalish, C. W., & Ge lman, S. A. (1992). On w o o den pillow s: Multiple c lassification and children’s

category-based inductions. Ch ild De v e lo p m e n t, 63, 1536–1557.

Lopez, A., Gelman, S. A., Guthe il, G., & Smith, E. E. (1992). The deve lopme nt o f c ategory-based

induction. Ch ild De v e lo p m e n t, 63, 1070–1090.

Mandler, J. M., & Mc Dono ugh, L. (1996). Drinking and driving don’t mix: Inductive gene ralization

in infanc y. Co g n itio n , 59, 307–335.

Morison, P., & Gardner, H. (1978). Dragons and dinosaurs: The child’s capacity to differentiate

fantasy from reality. Ch ild De v e lo p m e n t, 49, 642–648.

Piaget, J. (1929). Th e c h ild’s c o n c e p tio n o f th e w o rld . Savage, MD: Littlefield Adams.

Piaget, J. (1930). Th e c o n s tru c tio n o f re a lity in th e c h ild . Ne w York : Basic Boo ks.

Pre ntice , N. M., & Gordon, D. (1986). Santa Claus & the To oth Fairy for the Jew ish child and

parent. Jo u rn a l o f Ge n e tic Psy c h o lo gy , 148, 139–151.

Pre ntice , N. M., Mano se vitz, M., & Hubbs, L. (1978). Imaginary figures o f early c hildho o d: Santa

Claus, Easter Bunny, and the Tooth Fairy. Am e ric a n Jo u rn a l o f O rth o p s yc h ia try , 48, 618–628.

Ro se ngren, K. S., Kalish, C. W., Hickling, A. K., & Ge lman, S. A. (1994). Exp loring the relation

betw een pre schoo l children’s magical be liefs and c ausal thinking. Br itish Jo u rn a l o f

De v e lo p m e n tal Ps y c h o lo g y , 12, 69–82.

Samue ls, A., & Taylor, M. (1994). Children’s ability to distinguish fantasy e vents from real-life

eve nts. Br itish Jo u r n a l o f De v e lo p m e n ta l Ps y c h o lo g y , 12, 417–427.

Singe r, J. L. (1973). Th e ch ild ’s w o rld o f m a k e -b e lie v e : Exp e rim e n ta l s tu d ie s o f im a gin a tiv e

p la y . Ne w York: Ac ademic Press.

Singe r, D. G., & Singer, J. L. (1981) . Te le v is io n , im a g in a tio n , a n d ag g re s s io n : A s tu d y o f

p re s c h o o le rs . Hillsdale, NJ: Erlbaum.

Singe r, D. G., & Singer, J. L. (1990) . Th e h o u s e o f m a k e -b e lie v e : Ch ildre n ’s p lay a n d the

d e v e lo p in g im a g in a tio n . Cambridge, MA: Harvard University Press.

Sobel, D., & Lillard, A. (2001). The impac t of fantasy and ac tion on yo ung children’s

unde rstanding of pre tense . Br itish Jo u r n a l o f De v e lo p m e n tal Ps y c h o lo g y , 19, 85–98.

Spe lke, E., Guteil, G., & Van de Walle, G. (1995). The de velopme nt o f o bjec t pe rce ption. In S.

Ko sslyn & D. Osherso n (Eds.), Vis u a l c o gn itio n (2nd e d., Vol. 2, pp . 297–330). Cambridge,

MA: MIT Press.

Subbotsky, E. (1997). Ex planations o f unusual e vents: phe no menalistic c ausal judgeme nts in

childre n and adults. British Jo u rn a l o f De v e lo p m e n ta l Ps y c ho lo gy , 15, 13–36.

Taylor, M. (1999). Im a g in a r y c o m p a n io n s a n d th e c h ild re n w h o c re a te the m . New Yo rk:

Ox ford Unive rsity Press.

Taylor, B., & How ell, R. (1973). The ability of thre e -, four-, and five-year-o ld c hildren to distinguish

fantasy from reality. Th e Jo u rn a l o f Ge n e tic Ps y c h o lo g y , 122, 315–318.

Taylor, M., & Carlson, S. M. (1997). The relations betw ee n individual differe nce s in fantasy and

theo ry o f mind. Ch ild De v e lo p m e n t, 68, 436–455.

Taylor, M., Cartw right, B., & Carlson, S. (1993). A de velopme ntal inve stigation of imaginary

comp anions. De v e lo p m e n ta l Ps yc h o lo g y , 29, 276–2 85.

Vyse , S. (1997). B e lie v in g in m a g ic : th e p s yc h o lo g y o f s u p e rs titio n . Ox ford, UK: Ox ford

University Press.

We llman, H. M., & Este s, D. (1986). Early understanding of mental entities: a re e x amination o f

childhoo d re alism. Ch ild De v e lo p m e n t, 57, 910–923.

309Fantasy/reality distinction

We llman, H. M., & Gelman, S. A. (1998). Kno w ledge ac quisition in fo undational domains.

Ha n d b o o k o f Ch ild Ps y c h o lo g y (5th e d., Vo l. 2), pp. 523–574.

Wo o lley, J. D. (1997). Thinking abo ut fantasy: Are c hildren fundamentally different thinkers and

believers from adults? Ch ild De v e lo p m e n t, 68, 991–1011.

Wo o lley, J. D., & Phe lps, K. E. (1994). Yo ung c hildren’s p ractical reaso ning about imagination.

British Jo u rn a l o f De v e lo p m e n ta l Ps yc h o lo g y, 12, 53–67.

Wo o lley, J. D., & Wellman, H. M. (1990). Young children’s understanding o f realities, no n-

realities, and appe arances. Ch ild De v e lo p m e n t, 61, 946–961.

Wyve r, S. R., & Spe nc e, S. H. (1999). Play and diverge nt problem solving: Evidence of a re c iprocal

relationship. Ea rly Ed u c a tio n a n d De v e lo p m e n t, 10, 419–444.

Received 28 March 2002; revised version received 3 June 2003

Appendix: Questions asked in the property attribution task

Can X travel the w hole w o rld in o ne night?

Can someo ne touch X?

Can X be in different places at the same time?

Doe s X need to slee p some times?

Doe s X ge t o lder eve ry ye ar?

Can X get hurt?

Doe s X have parents?

Doe s X e at dinner w ith his/her family sometimes?

Can X have a pe t?

Doe s X dream some times?

Can X know w hat w e’re thinking?

Can X get his/her feelings hurt?

310 Tanya Sharon and Jacqueline D. Woolley