Co~ltition, 13 (1983) 103--l 28 3

Beliefsabout beliefs: Representationand constraircing

function of wrong bekfs in young children’sunderstanding
of deception

HEINZ WIMMER”
University of Salzburg

JOSEF PERNER
University of Sussex

‘A travelling salesman found himself spending the night at home with

his wife when one of his trips was unexpectedly cancelled. The two
of them were sound asleep, when in the middle of the night there
was a loud knock at the front door. The wife woke up with a start
and cried out, ‘Oh, my God! It’s my husband!* Whereupon the
husband leapt out from the bed, ran across the room and jumped
out the window.’
Schank and Abelson, 1977, p. 59.

Abstract

Understanding of another person’s wrong belief requires explicit represer;ta-

tion of the wrongness of this person’s belief in relation to one’s own knowl-
edge. Three to nine year old children’s understanding of two sketches was
tested. In each sketch subjects observed how a protagonist put an object
into a location x and then witnessed that in the absence of the protagonist
the object was transferred from x to location v. Since this transfer came as a
surprise they had to assume that the protagontst still believed that the object
was in x. Subjects had to indicate where the protagonist will look for the
object at his return. None of the 3-4-year old, 57% of 4-6-year old, and
86% of 6-g-year old children pointed correctly ,to location x in both
sketches. Of the many cases where J-6-year olds made an error the]; failed

*This project was financially supported by a research grant from the Stiftung Volkswagenwerk to
Professor Erwin Roth, University of Salzbuq. The international cooperation of the two authors was
supported by the Academic Link Interchange Scheme of The British Council. The authors wi-h to
thank Silvia Gruber for expert advice on procedures and for her help in data collection. Gratefully we
acknowledge the friendly cooperation of the ‘Amt fur Kindergarten und Horte der Salzburger
Landesregiexung (Frau Kfchdorfer)‘, of the ‘Stadtjugendamt Salzburg (Herr Artner)‘, and staff and
children in various kindergartens in Salzburg.
Requests for reprints should be directed to either Heinz Wimmen, Institut fur Psychologie der
Universitiit Salzburg, A-5020 Salzburg, Akademiestrasse 22, Austria, or Josef Peaner, Experimental
Psychology, University of Sussex, Brighton BN19QG, England.

OOlO-0277/83/010103-26/$08.10 0 Eisevier Sequoia/Printed in The Netherlands

104 H. Wirnmerand J. Perner

in only about 2O70 to remember the initial location correctl.v. As a test of

the stability of children’s representation of the protagonist’s wrong belief
the sketches continued with a statement about the protagonist S intention
to either deceive an antagonist or truthfur& inform a friend about the ob-
ject’s Ix&ion. independent of age, of those children who correctly thought
that the protagonist would search in x, 85% of the time they also correctly
thought that he would direct his antagonist to location y and his friend to
location x. This shows that once children can represent a person’s beliefs
they can constrain their interpretation of this person’s stated intentions
to the l*erson’s beliefs. In a more story-like situation another group of
children had to infer a deceptive plan from the depiction of a goal conflict
between two story characters and one character’s expedient utterance. At
the age of 4-S years children correctly judged this utterance as a lie only
28% of the time while S-6-year olds did so 94% of the time. These
results sugg~t that around the ages of 4 to 6 years the ability to repiasent
the re&ionship between two or more person’s epistemic srates emerges and
becomes firmly established.

Premack and Woodruff (1978) argued that one could demonstrate that
chimpanzees have a ‘theory of mind’, i.e., an ability to impute mental states
to themselves and others. In Pylyshyn’s (1978) explicatiori this means that
somebody who has a theory of mind does not only have a representation
about a state of affairs (x) and stands in certain relationships to these repre-
sentations (e.g., wanting x, believing x, etc.) but also represents these rela-
tionships explicitly. Pylyshyn refers to this ability as an ability for ‘meta-
representation’. This ability looms large in philosophical discussion as a
necessary precondition for moral responsibility, self-consciousness, and
social interaction (e.g., Dennett, 1978a).
It was thought that the best method of demonstrating an ability for meta-
representation in non-linguistic chimpanzees is to teach these animals to
deceive a competitor. Deceptive action is a good way of demonstrating the
presence of a theory of mind, because on the one hand, flexible application
of a deception strategy ought to have high adaptive value, while on the
other hand it requires the conceptualization of the deceived person’s wrong
belief as a subgoal in one’s planning strategy. Woodruff and Premack (1979)
succeeded in teaching deceptive pointing to 2 out of 4 chimpanzees but only
a.fter an extensive training effort of 5 months. This difficulty in teaching
chimpanzees to deceive is comparable to the difficulties encountered in
teaching them language. Thus it seems that the natural acquisition of lan-
guage and of a ‘theory of mind’ is a human characteristic.
 Beliefs about beliefs 105

Although the natural occurrence of deception is a convincing indicator of

meta-representational activity it is a very conservative measure because it
requires many additional complex planning steps. Since ch.ildren are capable
of language there is alternative access to their meta-representational ability.
One obvious indicator of this ability is children’s use of meta-representation-
al langiiage, i.e., their reference to people’s ‘wants’, ‘beliefs’, etc. In an
observational study Bretherton and Beeghly (in press) found that at the age
of 2 l/2 years the majority of children spontaneously used a substantial
vocabulary about perception, volition, major emotions, and knowledge. The
most popular of these words were used equally frequently for self and
others. Hood and Bloom (1979) reported that at 3 years children referred
quite frequently to their own and others’ intentions in response to ‘why’-
questions.
Experimental studies tested the correct use and understanding of these
words at an early age. For instance, Shultz et al. (1989) demonstrated that
3 to 5-year olds distinguished correctly betwe?en intended acts and unin-
tended behaviour, such as mistakes, reflexes, and passive movement. There is
also evidence that 4-year olds start to understand correctly some, although
not all aspects of ‘know’ and ‘guess’ (Johnson and Maratsos, 1977; Miscione
et al., 1978), of ‘remember’ and ‘forget’ (Johnson and Wellman, 1980) and
start to understand the presuppositions of these words (Macnamara et al.,
1976).
The studies cited above demonstrate that very young children already
explicitly represent and hence are able to verbalize the relation in which they
and others stand to their representation of propositional content.
A more complicated meta-reRresentationa1 problem arises when one has
to explicitly represent the difference between one’s own anl;l somebody
else’s relation to the same propositional conten.t. This problem arises when
one has to account for the lack of knowledge in another person. In a series
of studies on epistemie egocentrism subjects had information about some
event and also had information about the fact that another person had no
such information. Chandler and Greenspan (1972) and Flavell et al. (1968)
generated this difference in knowledge by introducing a late arriving by-
stander. Marvin ef al. (1976) and Mossler et al. (1976) used a different
technique based on selective presentation of sense specific information. For
instance, in one study children were shown a televisilon sketch in which a
boy was entering a house and asked his sister for a biscuit. Then they
watched the same sketch over again in the presence of their mother but with
the sound track turned off, so that it was clear that the mother did n& hear
the auditorily conveyed information. Sixty percent of 4-year and 85% of
S-year olds answered correctly that the mother knew that the boy entered
106 H. Wimmerand J. Perner

the house and that sh,e did not know that he asked his sister ror a biscuit.
These results show that from 4 years onwards children are atile to differen-
tiate between their own knowledge and the absence of this knol-,vledgein the
other person.
in contrast to these studies which investigated children’? ability to rep-
resent the absence of knowledge in another person the present series of ex-
periments investigated children’s competence in representing another per-
son’s definite belief which differs from what the subject knows to be true
This problem has not been investigated developmentally. It can, however,
be conveniently studied in the following formal paradigm which has been
outlined by Bennett (1978), Dennett (19786) and Harman (1978) indepen-
dently in their comments on the paper by Premack and Woodruff (1978):
The subject is aware that he/she and another person observe a certain state
of affairs x. Then, in the absence of the other person the subject witnesses
an unexpected change in the state of affairs from x to y. The subject now
knows that y is the case and also knows that the other person still believes
that x is the case.
In order to test subjects’ comprehension of the other person’s wrong be-
lief, stories like the following were constructed: A story character, Maxi,
puts chocolate into a cupboard x. In his absence his mother displaces the
chocolate from x into cupboard ‘y. Subjects have to indicate the box where
Maxi will look for the chocolate when he returns. Only when they are able
to represent Maxi’s wrong belief #(‘Chocolate is in x’) apart from what they
themselves know to be the case (‘Chocolate is in y’) will they be able to
point correctly to box x. This procedure tests whether subjects have an
explicit and defmite representation of the other’s wrong belief. Yet, there
is neither a problem in framing the test question by using mental verbs (e.g.,
‘What does Maxi believe?‘) nor are !;ubjects required to verbalize their knowl-
edge about other’s beliefs since a mere pointing gesture suffices.
The practical importance of representing another person’s wrong beliefs
consisis in the use of this representation as a frame of reference for inter-
preting or anticipating the other person’s actions. That is, the interpretations
and anticipations have to be constrained to the realm of the other person’s
beliefs. In order for such a constraining function to occur, the representation
of the other person’s beliefs has to be frnnly established. As a test of the
firmness of children’s representation of wrong beliefs, the ‘wrong belief
paradigm as outlined above was extended by describing Maxi as wanting to
tell something wrong or something true.
fn a competitive story version Maxi’s brother was introduced who, too,
wants the chocolate and asks Maxi where it is. Maxi (who wrongly believes
that the chocolate is in x) decides to tell his brother something wrong, Sub-
 Beliefs aiiout beliefs 107

jects are then asked ‘Where will Maxi tell his brother the chocolate is?‘. A
correct answer to this question depends on the correct interpretation of
Maxi’s intention to tell something wrong. Maxi’s message to his brother has
to be ‘wrong’ in relation to ‘Maxi’s wrong belief’. Ironically, this correct
interpretation can lead to the answer t+!at Maxi will try to mislead his
brother by telling him: ‘The chocolate is in y’, where it actually is. Hence
there is a temptation for subjects to make the typically ‘egocentric’, incor-
rect interpretation. Maxi’s message, whick he intends to be wrong, could
easily be construed as actually wrong, i.e.. It is interpreted as being ‘wrong’
in relation to the subject’s own knowledge.
In a cooperative story version Maxi asks his Grandpa for help and has the
explicit intention to tell the truth about the location of the chocolate.
Again, there are two possible interpretation; of ‘true’: the correct intergreta-
tion relates the ‘true’ statement to Maxi’s wrong belief, while the incorrect
one relates it to the subject’s own knowledge of the truth. The correct inter-
pretation leads to the answer that Maxi tells his Grandpa: ‘The chocolate is
in x’, where the subject knows that it is not.
Therefore, the correct interpretation of the instruction to make Maxi say
something wrong can lead to a statement that is actually true, while the
correct interpretation of the instruction to make him say something true
should lead to an utterance that is actually wrong. One can empathise with
the subject’s temptation to take the wrong interpretation: when asked to
make Maxi say something (which he thinks is) wrong or true, it is but natural
to make him say something that is actually wrong or true. This temptation
will meet little opposition if subjects are still uncertain about the distinction
between their own knowledge and the protagonist’s beliefs. Hence, we can
expect that correct interpretations will only occur once this distinction has
been firmly established in the child’s mind. The representation of Maxi’s
belief has to be firm enough so that the interpretation of his intention to say
something wrong or true can be constrained to the realm of his beliefs.

Experiment 1

Method

Subjects
Thirty-six children from several kindergartens and summer camps in Sal>:-
burg, Austria participated in this study. There were twelve subjects in eac.h
of the fohowing age groups: 4-5 years (4;0 to 4;7, 5 girls, 7 boys); 6-7
years (6;0 to 6;11, 6 girls, 6 boys), and 8-9 years (8;l to 9;3,4 girls, 8 boys).
108 R Wimtterand J. Perner

Materials

Three cassette containers and 3 large match boxes were used for the 3 hiding
locations. The three containers were glued equally spaced along the bottom
of one side of a 40 cm wide and 30 cm high polyesther wall. The match
boxes were glued onto the other side along the upper rim of this wall. The
wall was erected on a 80 cm by 40 cm platform. The colours of the contain-
ers were blue, green and red and yellow, red and white for the match boxes.
Fifteen cm paper cut-outs of a young boy and a young girl were used to rep
resent the absence and presence of the story protagonist.

hcedure and Design

Each child was told two stories. Each story existed in two versir>ns (Table 1).
In one version it was made clear that the protagonist wanted to cooperate
with another character in obtaining the hidden object; in the other version
the protagonist competed with an antagonist.
Story 1 is fully described in both its versions in Table 1. Story 2 had
essentially the same structure. It was set in a kindergarten where a little girl
hid her favorite book. When all children were on a walk a caretaker reshelved
the book. Upon returning from the walk the second character is introduced.
In the cooperative version it is the girl’s friend to whom shle offers to show
her book. And she: tells him where to find it. In the other version a child is
competing for the book and the little girl tries to mislead him.
Story 1 was always told before story 2. For half the subjects story 1 was
told in its cooperative version and story 2 in its competitive version. The
other half was told the complementary versions. The stories were tape
recorded and played to the subject while the experimenter carried out the
stage instructions. The test questions were asked by the experimenter.

ResuIts

Representing Wrong Beliefs

Table 2 shows the number of children who pointed correctly to location x in
response to the ‘Beliefquestion: e.g., ‘Where will Maxi look for the choco-
late?‘. All incorrect responses consisted of pointing to the actual location of
the chocolate (location y); there were no wrong responses of pointing to the
third location z. Table 2 reveals a strong age trend. A majority of 4-S-year
olds pointid wrongly to the actual location y while almost all 6-g-year olds
 Beliefs about beliefs 109

Table 1. Cover story I in its cooperative and competitive version including [stage
instructions]
--_ll -.-.---__I_ __-.__-_-__
Version
-- - _. .--
Cooperative Competitive
.--- ~-~- -
Is cene with match boxes fixed high [Scene with cas,srtte containers low on
on the wall 1 the wall1

[Boy-doll present; representing ?4axi waiting fox-his mother.)

“Mother returns from her shopping trip. She bought chocolate for a cake. Maxi may help
her put away the things. He asks her: ‘Where should I put the chocolate?’ ‘In the blue
cupboard’, says the mother.

‘Wait I’ll lift you up there, because you a~? t(;‘o small.’

Mother lifts him up.

Maxi puts the chocolate into the blue cupboard. [A toy chocolate is put into the ‘blue
match box.) Maxi remembers exactly where he put the chocolate so that he could come
back and get some later. He loves chocolate. Then he leaves for the playground. [The
boy doll is removed.] Mother starts to prepare the cake and takes the chocolate out of the
blue cupboard. She grates a bit into the dough and then she does nor put it back into the
blue but into thegreen cupboard. (Toy chocolate is thereby transferred from the blue to
the green match box.] Now she realizes that she forgot to buy eggs. So shegoes to hrr
neighbour for some eggs. There comes Maxi back from the playground, hungry, and he
wants to get some chocolate. (Boy doll reappears.] He still remembers where he had put
the chocolate.”
[‘BELIEF’question] ‘Where will Maxi look for the chocolate?’

(Subject has to indicate one of the 3 boxes. 1

“OK, there he’ll look, but he is too However, before Maxi gets a chance to get
small to reach up there. There comes at the chocolate his big, brother comes into
Granipa and Maxi says: ‘Dear Grandpa, the kitchen. He, too, is looking for the
pleas3 help me get the chocolate from chocolate. He asks Maxi where the choco-
the cupboard.‘,Grandpa asks: ‘Which late is. ‘Good grief’, thinks Maxi, ‘now big
cupboard?’ ” brother wants to eat up all the chocolate.
I will tell him something completely wrong
so that he won’t find it, for sure’.”
(‘UTTERAHCE’question1 ‘Where will Maxi say the choLwlate is?’

[The box indicated is opened.}

[‘REALITY’question (asked only if the box opened I?, found empty)] ‘Where is the
chocolate really?’
[‘MEMORY’question] ‘Do you remember where Maxi put the chocolate in the bcgin-
ning?’
110 H. WimmerandJ. Per’ner

Table 2. Numlw of bwbjects givingcorrectanwers to ‘Belief’questionin ExperimentI

--
Number of correct answers

Age (years) 2 -’ 1 0
-_
4-5 4 2 6
6-7 11 0 1
8-9 11 1 0

pointed to the correct location x. For a test of statistical reliability the data
of Table 2 were collapsed into a 2 X 2 contingency table (4-S-year olds
versus older children, and children who responded correctly on both occa-
sions versus children who gave wrong responses on at least one occasion):
Chi Square = 10.82, d.f. = 1, p < 0.0 1.
Wrong responses could not be due to a failure of remembering the actual
sequence of events because in 80% of the cases where children pointed to
location y in response to the ‘Belief-question they gave a correct answer to
the ‘Memory’question: e.g., ‘Do you remember where Maxi put the choco-
late in the beginning?‘). Conversely, correct answers to the ‘Belief-question
were not due to a failure to remember the actual location of the object,
because all children who answered the ‘Beliefquestion correctly, also gave
correct answers to the ‘Reality’question.

Lbnstruction of Utterances
Inspection of the data showed that there was no discernible effect caused by
the two stories. Thus, the data were collapsed. Table 3 separates the data for
those children who responded correctly to the ‘Belief-question from those
who responded ;ncorrectly. For deceitful utterances called for by the com-
petitive version the relevant data are in the first two rows of Table 3. The
last column shows that for children of all ages who answered the ‘Belief-
question correctly with location x, 23 were also able to construct an appro-
priately deceitful utterance by letting the protagonist direct their competitor
to either location y (20 of them) or z (3 of them). Only 5 gave incorrect
answers: Chi Square = 11S7, d.f: = 1, p < 0.001.
For truthful utterances in the cooperative version the analogous data are
in the third and fourth row of Table 3. The proportion of responses is quite
opposite to the one in the competitive version: Twenty-three subjects let
the protagonists direct their cooperating friends to location x, while only 4
let t.bem point ta location y: Chi Square = 13.37, d.J: = 1, p < 0.001.
 Beliefs about beliefs 111

Table3. Number of subjects who answered the ‘Belief’-qtrcstion with locations x

(correct) or y (incorrect) and gave the same or a different answer to the
‘Utterance‘question in Experiment I
.- --- -_____
Answer to story Answer to Age (years)
‘Belief question Version ‘Utterance’question
4-5 6-7 8-9 Total
-
Correct Competitive Same: x 1 3 1 5
(x1 Different: y or z 4 8 11 23
Cooperative Same: x 4 10 9 23
Different: y or z 1 1 2 4

Incorrect Competitive Same: y 7 1 8 8

(Y) DilTerent: x or z 0 0 0 0
Cooperative Same: y 5 1 0 6
Werent: x or z 2 0 13
- _ _.-- ___
NOW. In the competitive version the ~:rr::~:r.:zsoonse to the ‘Utterance’question consists
in giving a differentanswer than io the tklief’-.r;uestion, wGle in the cooperative version
the sume answer should be given. Since each child was toid two stories, one in its com-
petitive the other in its cooperative version, the numbers in the table add up to twice
the number of subjects in the experiment.

For the cooperative as well as the competitive story version Table 3 shows
that the proportion of correct answers to the ‘Utterance’question remained
roughly constant across ages for those children who responded correctly to
the ‘Belief question.
These results show that children at all ages who correctly ascribed to the
protagonist a wrong belief were also able to construct for the protagoniat
an utterance which is deceitful or truthful in reZation to the protagonist’s
belie5 Such a belief-dependent construction of utterances is a quite intricate
cognitive achievement in the present experimental situation. If the protago-
onist intended to point to the true location, then the subject had to point
for the protagonist to a location which is known to the subject as the actual-
ly false location. Conversely, if the protagonist intended to cheat and point
to a faZse location, then nearly all subjects pointed for the protagonist to the
actually true location. Many children were observed to react with an ironical
smile to the protagonist’s counterproductive deceptive effort.
The lower half of Table 3 presents the data for those subjects who re-
sponded incorrectly with the actual location y to the ‘Belief-question. Most
of these subjects responded with location y also to the ‘Utterance’-question
regardless of the truthful or deceitful intentions of the protagonists.
i 12 H. WimmerandJ. Pewter

Experiment 2

In this experiment two explanations are explored of why a high proportion

of the youngest subjects in Experiment 1 failed to correctly ascribe a wrong
belief to the protagonist.
A series of studies on impulsivity versus reflectivity in children (see review
by Kagan and Kogan, 1970) consistently showed for various task domains
that young children react much faster but, therefore, less accurately than
older children. Indeed, many of the 4-S-year old children in Experiment 1
gave the impression that they pointed automatically without much reflection
to the actual location y in response to the ‘Belief- and the ‘Wtterance’-
questions. To control such impulsive behaviour a ‘Stop and Think’ request
was introduced before the ‘Belief question.
Another explanation attributes the youngest subjects’ difficulty with the
‘Beliefquestion to a possible interference between mental representations.
Since the target object is merely displaced from a box of one colour to a
box of another colour, the description of the protagonist’s wrong belief, e.g.,
‘The chocolate is in the blue cupboard’, is fairly similar to the description of
the subject’s own knowledge that ‘The chocolate is in the green cupboard’.
Due to this similarity the representation of the protagonist’s belief may be
easily overwritten by the dominating description of the true state of affairs.
In order to reduce the similarity of descriptions a new condition was intro-
duced in which the target object was taken from the first cupboard (location
x) and then disappeared altogether from the scenes, i.e., the chocolate in one
story was used up, the picture book in the other story was taken home bj a
caretaker.

Method

Subjects
Ninety-two children from several kindergartens in Salzburg participated in
this study. There were 20 3-cl-year old (3;l to 3;9, equal number of boys
and girls), 42 4-S-year old (4;l to 4;9, 20 girls and 22 boys) and 30 S-6-
year old children (5 ;l to 5 ;9, 14 girls and 16 boys).

Material and hcedwe

The same two stories were used as in Experiment 1 with the following modi-
fications: for instance the story described in Table 1 was told up to the sen-
 Beliefs about beliefs 113

tence: ‘There comes Maxi back from the playground, hungry’. In the Stun-
durd Displacement condition the ‘Belief-question followed: ‘Where will
Maxi look for the chocolate ?‘. In the case of a correct response by the sub-
ject this question was followed by the ‘Reality’question: ‘Where is the choc-
olate really?‘, otherwise the ‘Memory’question was asked: ‘Do you remem-
ber where Maxi put the chocolate in the beginning?‘.
In the Stop-crnd-Think Displacement condition subjects were given a re-
minder before the ‘Belief-question: ‘Think carefully! What did Maxi do
before he went off to the playground? Now he wants to eat the chocolate.
Where will he look for the chocolate?‘.
In the Disappear condition the stories were altered at an earlier point.
Instead of using only part of the chocolate children were told: ‘The mother
grates all of it into the dough. Now there is no chocolate left. The chocolate
cake is for the grandmother who has her birthday today. When the cake is
ready the mother takes it and brings it over to Granny’s. [The chocolate is
removed from the scene and placed behind the wall.] Now Maxi returns
from the playground’. The ‘Belief-question followed: ‘What will Maxi do?‘.
If no answer or an unclear answer was forthcoming, a prompting question
was asked: ‘Now look, Maxi returns hungrily from the playground and wants
to eat the chocolate. Where will he look for the chocolate?’ Subsecguently
the ‘Reality’- or the ‘Memory’question was asked depending on subject’s
response.
The second story was modified accordingly. The caretaker did not re-
shelve the picture book but ;sok it home for redpair. Each child heard both
stories from a tape recorder. Their sequence was counterbalanced across age
groups and experimental conditions.

Results

Belief Questions
Table 4 shows the number of subjects who correctly attributed the wrong
belief to the story protagonist in both, only one, 3r in neither of the two
stories.
As in Experiment 1 subjects are bimodally d.istributed over these 3 catego-
ries (last row of Table 4). Under the assumption of a single peaked distribv-
tion the frequency in the middle should not be less than the smaller frequen-
cy of the two end categories, which can be reliably rejected: Chi Square =
25.13, d.f: = 1, B < 0.001.
114 H. WimmerandJ. Pemer

Table 4. Number of subjects who answeredcorrectly with locationx to ‘Belief’-

question (e.g., ‘WherewillMaxi look for the chocolate?~in Experiment2

Number correct

Age (years) Condition n 2 1 0

3-4 Stop-and-Think Displacement 10 0 0 10

Disappear 10 0 3 7
4-s Standard Displacement 14 6 1 7
Stop-and-Think Displacement 14 4 2 8
Disappear 14 11 0 3
S-6 Standard Displacement 10 s 0 s
Stop-and-Think Displacement 10 10 0 0
Disappear 10 10 0 0

Total 92 46 6 40
--

Five to six year olds found both of the new conditions helpful and scored
perfectly on the ‘Belief-question in both conditions. In the Standard Dis-
placement condition, however, only half of them responded correctly. The
frequency distribution in the Standard Displacement condition was reliably
different from the distributions in either one of the other two conditions
(Fisher’s Test: p < 0.025). This and all subsequent tests of statistical reliabil-
ity were carried out as in Experiment 1 by collapsing the number of subjects
who responded correctly only once with those who did not respond correct-
ly in either story.
The frequency distributions of the 4-S-year olds show a reliable differ-
ence across the 3 conditions: Chi Square = ‘7.43, d$ = 2, p < 0.05. From
Table 4 it is clear that this difference is due to better performance in the
Disappear condition only. There is no evidence that the Stop-and-Think con-
dition helped the 4-5-year olds subjects at all.
For the 3-4-year olds there is no reliable evidence that they can correctly
ascribe a wrong belief in any of the two experimental conditions.
At all ages children’s responses to the modified ‘Belief-question in the
Disappear condition merit closer inspection. In 50% of all cases 5-6-year
olds gave complete answers to the modified ‘Beliefquestion (‘What will
Maxi do?‘) by responding with location x. In the other 50% of cases they
initiaIly gave non-specific answers, e.g., ‘He will look for the chocolate’ but
responded specifically with location x to the subsequent prompting question
‘Where will Maxi look for it?’
 Beliefs about beliefs 115

Four to five year olds gave correct and specific answers to the ‘Belief-
question in 39.3% of all cases. In 50.0% of the times they responded non-
specifically at the beginning but correctly with location x to the prompting
question. In the remaining 10.7% they responded incorrectly to the prompt-
ing question by claiming that the protagonist would search for the object
behind the scenes.
None of the 3-4.year olds gave any answers to the first ‘Belief-question.
In response to the promp5ng question they responded correctly with loca-
tion x in only 15% of cases. whereas in all the remaining 85% of cases they
claimed that the protagonist would search behind the scenes.

Memory Controls
The ‘Memory’question, e.g., ‘Do you remember where Maxi put the choc-
olate in the beginning?’ was asked when subjects gave a wrong response to
the ‘Beliefquestion. Table 5 shows the percentage of incorrect answers to
the ‘Belief-question out of all trials and the percentage of correct answers
to the memory question out of all those cases where it was asked following
a wrong response to the ‘Beliefquestion.
The memory data for 4-6-year olds show that wrong answers tc:- the
‘Belief-question were followed in 83.7% by correct memory responses,
which replicates the finding of Experiment 1 where the respective percentage
was 80%. For the 3-4-year olds this percentage was substantially lower
(37.8%) which suggests that for the majority in this age group the task was
too complex. However, all their wrong responses consisted of the same
mistakes as the ones made by the older children: in the Stop-and-Think
condition they pointed to location y and in the Disappear condition they
answered with the location behind the scenes.
For all age groups all correct answers to the ‘Belief-question were fol-
lowed by correct answers to the ‘Reality’question, with the exception of
one 3 d-year old.
If one compares subjects’ performance on the ‘Memory’-questi& in the
Disappear condition with their performance in the Displacement conditions
a counterintuitive trend becomes apparent. The Disappear condition was
designed to test the interference hypothesis which gained some support from
the data on the ‘Belief-question. The description of the true state of affairs
was more similar to the description of the protagonist’s wrong belief in the
Displacement condition than in the Disappear condition, hence children
found it more difficult to keep the two delscriptions separate in the former
than in the Iatter condition. The same should hold true for memory of event
sequences. When the description of a previous position of an object is very
116 H. Wimmerard J. Perner

Table 5. Percentage of incorrect answers to the ‘BelieS_question(-Belief) and correct

answers to ‘Memory‘question (+Memory) as a percentage of those cases in
which wrong answers were given to the ‘Belief’-question in Experiment 2

Age and Question

3-4 years 4-5 years 5-6 years

-- --- --m
Condition -Belief + Memory -Belief + Memory -Belief + Memory
/-Belief /-Belief /-Belief

Standard Displacement - - 53.6 1oo.c 50.0 100.0

Stop-and-Think Displace- 100.0 50.0 64.3 66.7 0.0 -
ment
Dhappear 85.0 23.5 21.4 66.7 0.0 -.

similar to the description of the present position of this object (as in the
Displacement conditions) then it should be more difficult to keep the two
descriptions apart then when they are dissimilar (as in the Disappear condi-
tion). However, the data in Table 5 show no such trend. In fact, if there were
any difference at ali, it would be in the opposite direction: The 4-S-year
olds who answered the ‘Beliefquestion incorrectly in the Disappear condi-
tion gave only 66.7% correct answers to the ‘Memory’question as compared
to 81.8% for the same age group in the two Displacement conditions put to-
gether. Also, for the 3-4-year olds the results point into the same direction.

Experiment 3

Experiment 1 showed that almost all of those children who correctly as-
cribed a wrong belief to the story protagonist were able to construct truthful
or deceitful utterances in relation to this wrong belief. Those children, how-
ever, who were unable to ascribe wrong beliefs and wrongly thought that the
protagonist would search in the actual location y also pointed to location y
in respo.nse to the ‘Utterance’question. Thus, from among all 4-S-year olds
only l/3 were able to construct a deceitful utterance correctly. This low
performance, however, might lead to an underestimation of this age group’s
true competence, because the complexity involved in attempting to cope
with the protagonist’s wrong belief may have prevented them from following
the rest of the story with fuI1 capacity. Experiment 3 attempts to test young
children’s ability to construct deceitful utterances without the complexities
that arise from the representation of wrong beliefs. This can be achieved by
 Beliefs about beliefs 117

a minor modification of the stories used in Experiment 1. For instance in

Story 1, the mother takes the chocolate from hocation x and after having
used some of it, puts it back into location x (x-x Replacement condition),
while in the original version the chocolate was taken from location x and
then put into location y (x-y Displacement condition). If the low propor-
tion ,of successfully constructed deceitful utterances in Experiment 1 was
due lo belief difficulties, then a higher proponion can be expected in the
x-x :than in the x-y condition.

Method

Subjects
A group of 29 4-S-year old (4;1-4;9) and a group of 24 5--6-year old
children (5;1-5;9) served as subjects in this study. From the younger group
there were 5 girls and 7 boys in the x-x condition and 5 girls and 12 boys
in the x-y condition. That there was an unequal number of subjects in the
two conditions was caused by an erro- in allocating boys and girls to the
two conditions at the beginning of the experiment. In the older group there
was an equal number of girls and boys in each condition. All subjects at-
tended kindergartens in Salzburg, Austria.

Design
Each subject was told two stories, one in its cooperative the other in its
competitive version, The assignment of story to version and sequence of
presentation were counterbalanced. Both stories were appropriately adjusted
to the x-x condition. In all other respects the method of Experiment 1 was
followed.

Results

Inspection of the data revealed no perceptible difference in responses to the

two stories, hence the data were collapsed for stories. Responses to the
‘Belief-question in the x-y condition replicated the results of Experiment 1
and of the Standard Displacement condition in Experiment 2. Table 6 shows
the number of subjects in each age group who answered correctly with loca-
tion x in both, only one or neither of the two stories.
f 18 H. Wimmerand J. Perner

Table 6. Number of subjects giving correct answers to ‘Belief-question in x-y Dis-

plucement condition of Experiment 3

Number of Correct Answers

Age (r=s) 2 1 0
-
4-5 6 2 9
5-6 8 1 3
~-. -- -..-

The number of subjects who responded correctly in both stories was con-
trasted with the number of subjects who responded incorrectly in at least one
story. This contrast showed no reliable difference between age groups: Chi
Square = 1.66, d.f. = 1, p > 0.10. Of those cases in which wrong responses
were given to the ‘Belief-question 73.1% were followed by correct answers
to the ‘Memory’question. All correct answers to the ‘Belief-question were
followed by correct answers to the ‘Reality’question.
Inspection of responses to the ‘Utterance’question showed no difference
for the two age groups. For competitive story versions in the standard x-y
Displacement condition 18 out of 29 subjects (62.1%) were able to construct
a deceitful utterance. In the new x-x Replacement condition, not surprising-
ly, all subjects answered correctly with location x to the ‘Belief-question,
but only 11 out of 24 (45.8%) gave correct answers (location y or z) to the
‘Utterance’question. Counter to expectations, this percentage was percep
tibly not better than the percentage of children responding correctly in the
x-y Displacement condition (62.1%). A 5% confidence interva! showed that
in the most extreme case, allowed for by the interval, at least more than one
third of all subjects in the x-x condition would have given a wrong answer
(6 out of 24). This result suggests that the construction of a deceitful utter-
ance remains difficult for 4-6-year old children even when no complicated
assumptions about the wrongness of the deceiver’s beliefs are involved. This
result supports findings by Shultz and Cloghesy (1981) who found that not
before the age of 5 l/2 years did children show signs of deceptive behaviour
in a variant of the ‘Hide and Seek’ game.
For those subjects who responded correctl:y to the ‘Belief-question in
the x-y Replacement condition answers to the ‘Utterance’question repli-
cated the results of Experiment 1. In the competitive story version 1’1of
15 gave correct answers to the ‘Utterance’question. In the cooperative
vemion all 16 subjects who responded correctly to the ‘Belief-question
gave correct answers to the ‘Utterance’question. The response pattern to the
 Beliefs about beliefs 119

‘Utterance”question was reliably dependent on story version: Chi Square =

13.7,d.cf: = I,p <O.OOl.
The high proportion of correct utterance constructions by those 4-S-year
olds who gave a correct answer to the ‘Beliefquest;iou, suggests that there is
no age difference in the usage of the correctly constructed belief representa-
tion. In order to give an indication of this age irkdependence the relevant data
from Experiments 1 and 3 were put together for 4-6 versus 6-9-year old
subjects. Yet the resulting contingency table shows no reliable age difference
either for the competitive story version (Chi Square = 0.06, d.f. = 1, p >
0.90.) or for the cooperative version (Chi Square = 0.22, p > 0.70).

Experiment 4

Children’s answers to the ‘Utterance’-question in the x-x condition of

Experiment 3 showed that half of them had substantial problems construct-
ing a deceitful utterance even though the deceiver’s beliefs could have posed
no problems (i.e., x-x condition) and even though the protagonist’s deceit-
ful intention was sta.ted in such an obvious way that it almost entailed the
appropriate utterance (e.g., Maxi thinks: I will tell him something complete-
ly wrong SO that he won’t find it.) Studies on children’s conception of lying
(Piaget, 1932; Stern, 19 14; Wimmer et al., Reference note 1) showed that
children at this age have no difficulty whatsoever in judging an utterance as
false if it did not correspond to the real state of affairs. Thus, it seems that
the difficulty with the Replacement condition must be due to children’s
failure to understand deceptive plans, in this case, the protagonist’s plan to
induce a false belief in his opponent.
A difficulty in understanding deceptive plans, however, contrasts with
this age group’s preference for fairy tales, many of which are based on intri-
cate deceptive ploys, e.g., ‘Hansel and Gretel’ (Bruce and Newman, 1978).
Btihler (1977) characterized the period of 4 to 8 years as the ‘fairy-tale age’.
It could be the case that the younger children appreciate fairy tales not for
their deceptive ploy but for their archetypal figures and archetypal conflict
situations as emphasised by psycho-dynamic theorists (Bettelheim, 1976).
It has to be noticed that deceptive acts are quite differently described in
fairy tales than in Experiments 1 and 3, and therefore, quite different infer-
ential processes might be required. In the competitive story versions of
Experiments 1 and 3 subjects were explicitly told the goal of the protagonist
(e.g., Maxi w;ants to eat the chocolate), the protagonist’s anticipation of his
antagonist’s conflicting goal (e.g., Maxi thinks that his big brother wants to
eat the chocolate), and the protagoni,st’s explicit intention to deceive his
120 H. Wimmerand.I. Perner

antagonist (e.g., Maxi wants to tell him someth.ing completely wrong). From
these three bits of information subjects have to infer an appropriate utter-
ance which satisfies the protagonist’s intentions.. In contrast, in deceptive
fajl tales children are usually told the protagonist’s goal, the antagonist’s
goal, and the protagonist’s utterance. From this Ihe listener has to infer the
protagonist’s deceitful plan.
In Experiment 4 an attempt was made to assess 4-6-year old subjects’
ability to make such inferences. Two stories were constructed in which the
protagonist made a critical utterance whose truth was to be judged by the
subjects. This utterance was preceded by one of three contexts.
The deceptive context depicted the conflict between the protagonist’s and
antagonist’s goal. This conflict together with the critical utterance strongly
suggested a deceptive plan on the protagonist’s part. The inference of such
a plan is required in order to conectly judge the critical utterance as false.
In the deceptive context there was no additional information on which such
a judgement could be based.
Pilot work indicated that 4-year old children tended to judge the critical
utterance in the deceptive context as true. Thus an attempt was made to
facilitate their task by giving them an extremely deceptive confext. Factual
information was added to the deceptive context which contradicted the
protagonist’s critical utterance.
In order to control for the possibility that children might stereotypically
judge the critical utterance as false, an informative context preceded the
utterance in a control condition. Here, the critical utterance was made true
by the factual information given in the context.
If the younger children’s bad performance on the ‘Utterance’question in
Experiments 1 and 3 was due to a general inability to understand deceptive
plans then children can be expected to misjudge the truth of the critical
utterance in the deceptive context.

Method
Subjects
There were 24 4 1/‘t-year old (4;3--4;8, nine boys and 15girls) and 16 5 l/2-
year old children (5;4--5;8, eight boys and eight girls) from several kinder-
gartens in S&burg, Austria.

Table 7 shows the 2 stories that were used with their context variations and
test questions. All stories were presented by tape recorder. The critical utter-
 Beliefs about beliefs 121

Table 7. Ckitical uttemnces in two stories in a Deceptive, (Extremely Deceptive) and

Informative context
-
STORY 1
Setting
Nancy lives in a large house. Behind the house is a playground with a sand box, a slide,
a climbing frame and a swing.
Deceptive Context Informdive Context

(Extremely Deceptive: One day Nancy walks One day Nancy walks to the playground
to the playground and meets the mother of and meets the mother of Thomas.
Thomas. Thomas’ mother says to her: Thomas’ mother says to her:
“Nancy, if you meet Thomas, please tell him, “Nancy, if you run into Thomas, please
that he should stay longer on the playground telI him that I need him and that he
today. I will come over myself later.“) should come home.”
Nancy wants to use the swing today. On the On the playground Nancy sees that
playground she sees that Thomas is using the Thomas is using the swing. She says
swing. Thomas never gets off the swing, but to Thomas:
Nancy wants to use it immediately, she says
to Thomas:
Critical Utterance
“Hallo Thomas, you’ve got to go home, your mom needs you.”
Test Questions
(a) Is it really true that Thomas has to go home?
(b) Why did Nancy say to Thomas that he has to go home? (In case of answer: “Because
mother needs him” an additional question was asked (b’): “Why did Nancy say to
Thomas that his mother needs him?“)

STORY 2
Deceptive Context Inform0 tive Context

(Extremely Deceptive: When Markus plays Markus hurt his foot playing football.
football Markus hurts his foot quite often. His foot hurts badly. Now he sits in his
Today he was playing football, but he did room and plays with his LEGO-blocks.
take great care, so that he didn’t hurt his foot He just started to build a talI tower.
once.) Markus is now playing in hi; room with His friend Rudi enters the room.
LEGO-blocks. He just started to build a tall Rudi says: “Markus, come and play
tower. His mother enters and says: “Markus football.” Markus answers:
could you, please, hop over to the store and
get me some milk”. Markus, howeva, wants
to finish his tower and says to his mother:
Critical Utterance
“I can’t go, my foot hurts. 1 hurt my foot playing football.”
Test Questions
(a) Is it reaIIy true that Markus hurt his foot?
(b) Why dkl Mrirkus say that his foot hurts? (In case of answer: “Because he hurt his
foot”, an additional question followed (b’): Why did Markus saq he hurt his foot?
122 Il. WimmerandJ. Ferner

ante was recorded only once in a neutral voice and spliced into each of the
3 context variations.

Design and Procedure

Each subject was told each story with the same type of context. Eight sub-
jects in each age group received each story with the deceptive context,
another eight with the informative context. A third group of only 4 l/2-year
olds was given the version with the extremely deceptive context. Each group
was subdivided into subjects who were told story 1 first and then story 2
and subjects who were presented with the two stories in the opposite order.
Also the order of the two test questions (a) and (b) was counterbalanced
within each experimental group.

Results

The best indicator that the critical utterance was interpreted in relation to
the deceptive or informative context are the correct answers to test question
(a) about the truth of the utterance. Appropriate interpretation should lead
to a ‘false’ judgment in the deceptive context and a ‘true’ one in the infor-
mative context.
Table 8 shows that there was no difference betwei?n the two stories. In
the informative context subjects in both age group: tended to give appro-
priate judgments. Children in the younger group terjded to treat the critical
utterance always as true even in the deceptive aiad extremely deceptive
contexts. However, almost all 5 l/2-year olds gave a context sensitive inter-
pretation: Fisher’s Exact Probability Test showed the difference in re-
sponse patterns between deceptive and informative context as statistically
reliable: p < 0.005 for both stories. Also, in the deceptive context the
response pattern was reliably different for the two age groups: p < 0.025.
This result obtained further support from answers to test question (b) or
the additional question (b’). Context sensitive answers should include at least
one reference to the ‘real’ goal of the person who made the critical utterance
(deceiver) in the deceptive context (e.g., ‘Nancy wants to use the swing’),
and in the informative context a reference to the source of the information
given in the utterance (e.g., ‘Thomas has to go home because his mother said
so’ or ‘...because his metier needs him’.). Regardless of age, responses to this
test question were in 88% of all cases appropriate in the informative context.
In the deceptive context 4 l/2-year olds produced only 50% appropriate
reSPonSesYwhereas, 5 l/Z&year olds gave 94% appropriate answers. The dif-
 Beliefs about beliefs 123

Table 8. Number of subjectsjudging the criticalutterance as tnre or false

Age and Context

4-S years 5-6 years

Story Answers Extreme Deceptive Informati-e Deceptive Informative

1 (Nancy) ‘True’ 6 5 8 0 6
‘False’ 2 3 0 8 2
2 (Markus) ‘True’ 6 6 7 1 8
‘False’ 2 2 1 7 0

erence in response pattern between contexts was statistically reliabHe for

5 l/2-year old children: p < 0.025 for both stories.
The results from the 4 l/2-year olds extend the finding of Experiment 1
and 3 that children at this age have a general difficulty in comprehending
deceptive ploys. That these children tended to misjudge the truth of the
critical utterance even in the extremely deceptive context came as a surprise,
since the falsity of the utterance could have been recognized directly by its
blatent contradiction with the factual information given in this context
condition. However, studies by Markman (1979) and Wimmer (1979)
showed that even older children had difficulty detecting the contradiction
between story parts even when they were explicitly asked to do so.
The impressive performance by the older, 5 l/2-year old children indicates
that children at this age may find it easier to infer a deceptive plan from the
story character’s conflicting goals and the protagonist’s utterance than to
construct a deceitful utterance from the conflicting goals and the protag-
onist’s deceptive intentions. Yet, the gap between acquisition ages for these
skills cannot be very large since the 6-7-year old subjects in Experiment 1
were quite able to construct deceitful utterances.

General discussion

Representing wrong beliefs requires the construction of two different models

of the world (cf., Johnson-Laird, 1980) and the explicit representation of
the falseness relation between propositions in one model and the correspond-
ing propositions in the other model. Within the experimental situation that
was explored children of 6 years and older were perfectly able to cope with
these representational complexities. This finding stands in a potentially
124 H. Wimmw and J. Perner

interesting contrast with findings about adults’ logicalreasoningcapacity.

Johru:mn-l&dand Steedman (1978) reported that adults found it difficult
to generate valid conclusions for a syllogism if this required two models or
more, In the context of false beliefs six-year olds have no difficulty in rep
resenting two models and using this representation to construct appropriate-
ly deceitful utterances in relation to these models. Hence the problem which
adults encounter with syllogisms which require more than a single model
may not only be due to the number of models involved. Part of the difficul-
ty could be caused by the factual form in which syllogisms are stated. Factu-
al information usually enables one to specify a single model in more detail,
whereas the main purpose of referring to people’s beliefs is to direct &ten-
tion to the possibility that people entertain different models.
Four to six-year old children also showed signs of being able to represent
wrong beliefs but were susceptible to minor modifications in the task. When
the description of the correct belief was dissimilar from the wrong belief, as
in the Disappear condition of Experiment 2, already 78% of the 4-5-year
olds constructed the wrong beliefs correctly. The performance of this age
group contrasted sharply to the performance by 3-4-year old children. Only
15% of answers by the youngest age group were correct. This abrupt devel-
opment step at the age of 4 was also documented by Marvin et al. (1976),
and Mossler et al. (1976). Marvin et al., e.g., found that 80% of 4-5-year
olds represented correctly another person’s absence of knowledge whereas
only 5% of 3-4year olds were able to do so. The close correspondence
between their results and the results from the Disappear condition of Exper-
iment 2 is remarkable in two respects. First, the representation of another
person’s lack of knowledge does not require a definite representation of a
particular proposition. Hence, the danger of interference between the two
representations is even less than in the Disappear condition. Yet the 3-40
year olds were not able to represent another person’s absence of knowledge
any better than another person’s wrong belief. Second, it could be argued
that the real life situations used by Marvin et al. (1976) and Mossler et al.
(1976) allow children to show their full competence better than in the story
formats used in the present series of experiments. Thus it is again important
that their 3--4-year olds did no better in the real life situation than in the
story format situation.
children’s answers to the ‘Utterance’question showed that independently
of age nearly ail those who were able to construe another person’s wrong
belief were also able to construct a deceitful as well as a truthful utterance
for this person. Correct construction of Maxi’s utterance requires a rather
mmplex representation of the cont6nts of his mind. Apart from this addi-
tional complexity the boundary around Maxi’s mental contents has to be
 Beliefs about beliefs 125

firmly established so that it con,.. +zitzs the interpretation of the ambigous

instruction to make him say something true orfb!z. The constraints have to
be SO strong that subjects’ are not susceptible to the obvious temptation to
make his utterance just ‘true’ or ‘false’, but that they make it ‘true’ or ‘false’
in reiuhn to Maxi’s wrong belief:
In all our experiments children had to infer a protagonist’s beliefs from a
sequence of events. The results showed that they become able to demon-
strate this inferential skill between ages 4 to 6 years. However, from the
younger children’s failure to show any sign of such a skill one cannot direct-
ly conclude that they are incapable of drawing inferences as Bryant and
Trabasso (1971) have cautioned us about children’s ability to draw transitive
inferences. These authors were able to show that if one assures that children
remember the facts, then all 4-year olds can draw the correct inferences. In
contrast to their finding, in the present series of experiments 80% of 4 to
6-year olds who failed to infer the protagonist’s belief correctly remembered
the facts. Hence, their failure to infer beliefs cannot be caused by lapses of
memory.
Pascual-Leone (1970) and CW, (1978) proposed that improved perfor-
mance on cognitive tasks with age .;an be explained by a stepwise increase in
central processing capacity.. One c’ould argue that such an increase is at the
heart of the observed development. At the age of 4 and younger the child
has at best enough capacity to remember the sequence of actual events,
whereas at 6 years the capacity has increased sufficiently so that the protag-
onist’s wrong belief can be represented in addition to the real events. The
plausibility of this account is, however, thwarted by our finding that 4 to
6-year olds become able to represent not only a wrong belief but also the
protagonist’s intentions and actions in relation to this belief If mental devel-
opment depended mainly on an increase in processing capacity one would
expect that this additional complication in the story should not be under-
stood for at least another two years.
Case (1978) emphasises a particular aspect of Pascual-Leone’s original
proposal. More complex tasks can be solved without the necessary increase
in capacity if experience with the strategy in question leads to automaticity
in the basic operations. In the context of belief inferences this would mean
that the protagonist’s intentions and actions can be represented in addition
to his wrong belief if the strategies for inferring his wrong belief had become
automatic. Yet, for this automation to occur one would expect that some
years of practice were required. However, in the pres’ent data there is no
indication of such a delay which would be necessary for an increase in men-
tal capacity or an automation of strategies to take place.
126 H. Wimmer and J. Pemer

In summary it seems, therefore, that t.he emergence of children’s ability

to understand another person’s beliefs ar,:d how this person will react on the
basis of these beliefs and their understanding of deception is not a mere side
effect of an increase in memory and central processing capacity. Rather, a
novel cognitive skill seems to emerge within the period of 4 to 6 years. Chil-
dren acquire the ability to represent wrong beliefs and to construct a deceit-
ful or truthful utterance relative to a person’s wrong beliefs. Within this
period several other relateId abilities also emerge: children start to under-
stand another peIson’a absence of knowledge (Marvin et al., 1976; Mossler
et al., 1976). They become able to construct a deceitful ,iatterance (x-x
Replacement condition of Experiment 3), and to infer a deceptive plan from
a critical utterance in the context of conflicting goals in Experiment 4.
Within this age span they also start to understand the competitive nature of
the ‘Hide and Seek’ game (Shultz and Cloghesy, 198 1). Correct performance
in all these tasks depends on correct representation of two different or even
conflicting epistemic states. The convergence of findings from such a variety
of different tasks suggests that around the ages of 4 to 6 years the ability to
represent the relationship between two or more persons” epistemic states
emerges and becomes firmly established within these two years.

References

Benett, J. (1978) Some remarksabout concepts. Behuv. BrainSci., I, 557-560.

Bettelheim, B. (19761 The Use of Enchuntmenr: The Meuningand Importance of Fairy Tcrles.Lon-
don, Thames & Hudson.
Bretherton, I. and Beeghly, M. (in press) Talking about internal states: the acquisition of an explicit
theory of mind. Devt$. Psyclrol.
Bruce, B. and Newman, D. (1978) Interacting plans. Cog. Sci., 2, 195-233.
Bryant, P. E. and Trabasso, T. (1971) Transitive inference and memory in young children. Nature,
232,456~458.
B(ihler, C. (1977) Das M&hen und die Phantasie des Kindes. Berlin, Springer.(Originally published,
1918).
Case, R. (1978) Intellectual development from birth to adulthood: A neo-Piagctian interpretation. In
R. S. Siegler (ed.), Children’sthinking: Whutdevelops? Hillsdale, NJ, Erlbaum.
Chandler, M. J. and Greenspan, S. (1972) Ersatz egocentrism: A reply to H. Borke. Devel. Psychol., 7,
104-106.
Dennett, D. (1978~) BroinsZoms. Montgomery, Vt., Bradford Books.
Dennett, D. C. (19786) Beliefs about beliefs. Behov. Bruin Sci., I, 568-570.
Bennett, D. (1978) Brainstorms.Montgomery, Vt., BradfordBooks.
Dermett, D. C. (1978) Beliefs about beliefs. Behav.B&n Sci.. 1,568-570.
Flavell, 1. H., Botkin, P. T., Fry, C. L., Wright, J. W. and Jarvis, P. E. (1968) The Development of
Role-taking and Communication Skills in Children. New York, Wiley.
 Beliefs about beliefs 127

Harman, G. (1978) Studying the chimpanzee’s theory of mind. Belzav. Bruin Sci., I, 515-526.
Hood, L. and Bloom, L. (1979) What, when, and how about why: A longitudinal study of early ek-
preSSiOn Of causality. Monographs of the Society for Research in Child Development, Serial
No. 181.
Johnson, C. N. and Maratsos, M. P. (1977) Early ccmprehension of mental verbs: think and know.
ChikdDevel., 48.1743-1747.
Johnson, C. IV. and Wellman, 21. M. (1980) Children’s developing understanding of mental verbs:
Remember, know and guess. Child Devel., 51. 1095-l 102.
Johnson-Laird, P. N. (1980) Mental models in cognitive science. cog. Sci., 4, 7 I- 1 I 5.
Johnson-Laird, P. N. and Stcedman, M. .I. (1978) The psychology of syllogisms. Cog. Psych&, IO,
64-99.
Kagan, J. and Kogan, N. (1970) lndividuality and cognitive performance. Pp. 1273-1366 in P. H.
Mussen ted.), Carmichaels’s Mum.& of Child Psychology. (Third Edition), Vol. 1. New York,
Wiley.
Macnamara, J., Baker, E. and Olson, C. (1976) Four-year-olds’ understanding of pretend, forget, and
know: Evidence for propositional operations. Child Dewl,, 47,62-70.
Markman, E. M. (1979) Realizing that you don’t understand: Elementary school children’s awareness
of inconsistencies. Child Devel.. 50.643-655.
Marvin, R. S., Greenberg, M. T. and Mossier, D. G. (1976) The early development of conceptual
perspective taking: Distinguishing among multiple perspectives. Child Devel., 4 7, 5 11-S 14.
Miscione, J. L., Marvin, R. S., O’Brien. R. G. and Greenberg, M. T. (1978) A developmental study oi
preschool children’s understanding of the words ‘know’ alid ‘guess’. Child Devel., 49, 1107-
1113.
Mossler, D. G., Marvin, R. S. and Greentlerg, M. T. (1976) Conceptual perspective taking in 2- to 6-
year-old children. Devei. Psychoi., 12.85-86.
Pascual-Leone, J. (1970) A mathematical model for the transition rule in Piaget’s developmental
stages. Acta Psychol., 32, 301-345.
Piaget, J. (1965) The Moral Judgment of tire Child. New York, Free Press. (Originally published,
1932).
Premack, D. and Woodruff, G. (1978) Does the chimpanzee have a theory of mind? Behav. Brain Sci..
I, 515-526.
Pylyshyn, 2. W. (1978) When is attribution of beliefs justified? Behav. Bruin Sci., 1,592-593.
Schank, R. and Abelson, R. (1977) Scr,pts, Plans, Goals, and Understanding. Hillsdale, NJ, Erlbaum.
Shultz, T. R. and Cloghesy, K. (1981) Development of re.ursive awareness of intention. Devel. Ps_v-
chol., 17,465-471.
Shultz, T. R., Wells, D. and Sarda, M. (1980) The development of the ability to distinguish intended
actions from r&takes, reflexes, and passive movements. Brit. J. sot. clin. Psyckol., 19, 3Ol-
310.
Stern, W. (1971) Psychologie der friihen Kindheit. Darmstadt, Wissenschaftliche Buchge:seUschaft.
(Originally published, 1914)
Wellman, H. M. and Johnson, C. N. (1979) Understanding of mental processes: A developmental study
of ‘remember’ and ‘forget’. Child Devel., 50.79-88.
Wimmer, H. (1979) Processing of script deviations by young children. Discourse fiocesses, 2, 301-
310.
Woodruff, G. and Premack, D. (1979) Intentional communication in the chimpanzee: The develop
ment of deception. Cog., 7, 333-362.
128 H. Wimmerund J. Pemer

Reference Notes

I Wimmer, H., Gruber, S. and Pemer, J. Being wren;: in good faith: Children’s concept of kying.
(Unpublished manuscript, University of Salzburg, Salzburg, Austria, July 1981.)

Comprcndre que cc que croit un tiers est erron& rcquiert une rcprEsentation explicit& de cette fausse
croyance en relation avec son savoir propre.
On a test4 la comprihension de deux sketches par des enfants de 3 i 9 ans. Dans chacun dcs
sketches les sujets obscrvent un protagoniste placer un objet dans un lieu ‘x’, puis sont tt!moins du
transfert de cet E;bjet de ‘x’ en ‘y’ en Sabsence du protagonistc. Ce transfcrt doit causer une surprise
chez le protagoniste dont on assume qu’il croit que I’objet se trouve toujours en ‘x’. Lcs sujets doivcnt
dire oi le protagoniste va chercher I’objet. Aucun 3-4 ans n’indiquc correctcmcnt le lieu ‘x’, 57%
des 4-6 ans et 86% des 6-9 ans le font. Parmi les nombreuses crreurs dcs 4- 6 ans seules 20% sont
attribuables 1 une incapaciti de se souvenir du lieu ‘x’. Pour tester la stabi!iti? de la repr6sentation de
la croyance error&, on dit que le protagoniste a I’intcntion soit de trompcr un advcrsaire soit d’in-
former un ami sur ie lieu oti se trouve i’objet. Indipendamment de leur Qe, les enfants ayant donni
des riponses correctes disent correctement dans 85% des cas que le protagoniste conduirait I’advcrsaire
en ‘y’ et I’ami en ‘x’. Lorsque les enfants se reprksentent les croyanccs d’une personne, ils peuvent
faire dCpcndre leurs interp&ations des intentions exprimecs par celles-ci g partir dc scs croyances.
Dans une situation de type histoire, un autre groupe d’enfants doit inf<rcr un essai de tromperie
& partir de la rep&entation d’un but conflictuel entre deux des personnages dc I’dnon& tactique
d’un des personnagcs. A 4-5 ans les enfants ne jugent correctement cet how6 comme mcnsonger
que dans 28% des cas alors qu’i 5-6 on a 94% de reporrses correctes. Les resultats indiquent que vers
4-6 ans la capacitC de rep&enter une relation entre les itats Qpistimiqucsde deux personnes ou plus
kmergeet se confirme.