1. family and parenting
2. children

OF
4
5
Theory of mind finds its Piagetian perspective:
why alternative naming comes with
understanding belief
Josef Perner a,∗ , Sandra Stummer a , Manuel Sprung a ,
Martin Doherty b
5
6
a
7
8
9
TE
3
DP
RO
Cognitive Development 103 (2002) 1–22
Department of Psychology, University of Salzburg, Hellbrunnerstrasse 34,
A-5020 Salzburg, Austria
b University of Stirling, Stirling, Scotland, UK
Abstract
11
26
In the last of a series of experiments 48 3–5-year old children were tested on an alternative naming game with “synonyms,” e.g., if puppet calls the depicted item a “rabbit” the
child has to call it a “bunny,” or the child has to judge puppet’s performance when roles
are reversed. The game was also played with categories (rabbit–animal), name/colour
(rabbit–black), colour/colour (black–white), and part/part (head–tail). The younger children (≤3.5 years) had severe problems with “synonyms” and categories (alternative names
for the items, <10% correct), but not with names and colours, only colours, or only parts
(>80% correct). Children’s increasing success with age on the alternative names tasks was
closely paralleled (.53 ≤ r ≤ .72) by their mastery of the false belief task in which they
had to predict that a mistaken story character would look for a desired object in the wrong
location. For explaining the synchrony between alternative naming and understanding
false belief we draw on the Piagetian idea that children come to represent perspective at
some point in their development. To apply this idea to the alternative naming game we
draw on the philosophical discussion about sortals (terms that specify what sort of object
something is) creating perspective differences.
© 2002 Published by Elsevier Science Inc.
27
Keywords: Perspective; Preamble; Egocentrism
14
15
16
17
18
19
20
21
22
23
24
25
RR
13
∗
CO
12
EC
10
1
2
UN
Corresponding author. Tel.: +43-662-8044-5124; fax: +43-662-6389-5124.
E-mail address: [email protected] (J. Perner).
0885-2014/02/$ – see front matter © 2002 Published by Elsevier Science Inc.
PII: S 0 8 8 5 - 2 0 1 4 ( 0 2 ) 0 0 1 2 7 - 2
J. Perner et al. / Cognitive Development 103 (2002) 1–22
OF
2
1. A Piagetian preamble
29
55
Perspective played a central role for Piaget. Intellectual development consisted for him of overcoming one’s egocentrism by progressively decentering
from how the world appears within ones point of view (perspective) to an objective, perspective-independent understanding. For instance, the young infant has
to learn that objects do not go out of existence as they disappear from one’s view.
Piaget and Inhelder (1948/1956) also investigated the child’s ability to represent
(understand) visual perspective. We make use of the notion of representing (understanding) perspective to explain an otherwise mysterious developmental relation
between children’s understanding of false belief and their appreciation of alternative labels for objects in a naming game.
Young children’s problems appear to stem from a mutual exclusivity (that
each thing can only have one name) or single identity constraint (Flavell, 1988;
Markman, 1989). In order to find a common denominator for understanding multiple identities and understanding false belief we propose that both rest on an
understanding of perspective. This proposal is in the spirit of Carey (1985) as a
compromise between Piaget’s formulations of domain general logical constraints
defining developmental stages and modularism, where each domain of knowledge
has its own developmental schedule (e.g., Leslie, 1994, for theory of mind).
The difficult step is to show why application of alternative labels to objects creates a perspective difference. We achieve a workable definition of “perspective”
with a constructivist conceptual analysis in Piaget’s spirit of decentering. Philosophical analyses of alternative object individuation with different labels suggest
the involvement of different perspectives. This allows us to show that both the false
belief task and the alternative naming game require an understanding of perspective. At the end we discuss how understanding perspective relates to children’s
mutual exclusivity bias and to classic Piagetian tasks like the class inclusion problem. We start with the empirical part.
56
2. Alternative naming and false belief
57
2.1. The original “synonyms” and false belief finding
58
Doherty (1994) and Doherty and Perner (1998) reported a surprisingly strong
and robust developmental synchrony between children’s understanding of false
belief and their ability to master a “synonyms” task. In the false belief task children
are told about Max, who puts his chocolate bar into one location (A) and leaves.
In his absence the chocolate is unexpectedly transferred to another location (B).
He returns hungry for his chocolate. Children are asked where he will look for
the chocolate. Typically young 3-year-olds answer incorrectly that he will look
in location B where the chocolate bar actually is. After 4 years most children
give the correct answer that he will look in A (which is empty but where he
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
59
60
61
62
63
64
65
66
TE
34
EC
33
RR
32
CO
31
UN
30
DP
RO
28
OF
J. Perner et al. / Cognitive Development 103 (2002) 1–22
3
98
2.2. Competing explanations
99
Doherty (1994) had designed the “synonyms” task as test of the theory that
children’s mastery of the false belief test reflects a deeper understanding of mental
states as representations (Doherty, 1994; Flavell, 1988; Forguson & Gopnik, 1988;
Perner, 1988, 1991, dubbed it RUM “Representational Understanding of Mind”). It
implies that children apply their newly gained understanding to beliefs but also to
other, non-mental representations. Zaitchik (1990) had explored this expectation
in the realm of pictorial representations. The “synonyms” task was designed to
check this implication for understanding linguistic expressions as representations.
It uses the fact that language as a representational medium can have different representational forms (words) to represent the same content (meaning). To succeed
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
100
101
102
103
104
105
106
107
108
TE
72
EC
71
RR
70
CO
69
UN
68
DP
RO
97
thinks it is). This happens at the same age at which they pass the “synonyms”
task.
In the “synonyms” task children were first tested for their knowledge of word
pairs that come close to being synonyms, e.g., bunny–rabbit, lady–woman,
television–TV, etc. On different occasions children were shown the same object
among three distractors and asked, “Which one is the bunny?” and later, “Which
one is the rabbit?” In the task proper a speaking puppet named an object, e.g., rabbit, and children had to “help” the puppet provide an alternative name, bunny. The
task was first demonstrated with feedback on three modelling pairs. The experimenter provided the answers if the child could not, and explained that the answer,
e.g., bunny, was correct because the object had two names, bunny and rabbit. Then
five test pairs were presented. The pairs were presented twice, once with the puppet
providing one of the synonyms, then the other synonym. This was to ensure that
children could not succeed by using their habitual term. Younger children tended
to repeat puppet’s expression, while most of the older children correctly used the
alternative word on both trials. Performance on the “synonyms” task and the false
belief task correlated strongly (r =≥ .65) even with children’s age and verbal intelligence partialled out (pr ≥ .60). Because children had to produce a suitable alternative expression, this was called the production version of the “synonyms” task.
The paradigm was also played with reversed roles. The child named the item
first (e.g., “a rabbit”) and the puppet then had to produce the alternative expression.
The child was to judge whether the puppet had conformed to the instructions in
each of three trial types in a random sequence (judgement version). The puppet
either (1) correctly used the alternative description (“a bunny,” correct judgement:
“yes, puppet said the right thing”), or (2) repeated the child’s label (“a rabbit,”
correct judgement: “no”), or (3) said something wrong (“an elephant,” correct
judgement: “no”). Again correctness of judgements was strongly correlated with
passing the false belief task (r ≥ .82, pr ≥ .70). These correlations between the
false belief task and the judgement version of the synonyms task have since been
replicated by Doherty (1998), Clements, Heider, Brooks, and Garnham (1998),
and Garnham, Brooks, Garnham, and Ostenfeld (2000).
67
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
DP
RO
114
TE
113
EC
112
RR
111
on the task children have to pay attention to the sameness of meaning of synonyms
(they can’t allow a name that is not a synonym) and at the same time ensure that
the second label does not have the same form as the first one, i.e., be a synonym
and not a repetition of the first label. The data bore out strongly the expectation
that the “synonyms” task and the false belief task are mastered at the same age as
predicted by the theory that they both require an understanding of representation.
An alternative explanation for the data by Doherty and Perner (1998) can be
derived from a speculation by Flavell (1988, pp. 245–246) and Markman (1989,
p. 212). They suggested that children’s tendency to assume that each object only
has one ‘name’ may have the same origins as children’s inability to understand the
appearance–reality distinction (Flavell, Flavell, & Green, 1983) and false belief
(Wimmer & Perner, 1983). Mutual exclusivity has been introduced as a principle that helps infants and children map new words to objects (Carey & Bartlett,
1978; Dockrell, 1981; Dockrell & Campbell, 1986; Markman & Wachtel, 1988;
Merriman & Bowman, 1989). Although, in the “synonyms” tasks children are not
learning new words but have to operate with ostensively familiar words, a kind of
conscious mutual exclusivity constraint could be evoked when children are directly
faced with the fact that more than one label exists for an object.
An obvious implication of this “ME hypothesis” is that children should not only
have problems with “synonyms” but with any two alternative labels for the same
object, e.g., basic and superordinate category labels: rabbit–animal. In contrast,
the representational-understanding-of-mind (RUM) hypothesis would predict that
this task should be substantially easier than the synonyms task, because “this is a
rabbit” and “this is an animal” have clearly different meanings and the task can be
mastered without understanding “rabbit” and “animal” as pure variations of word
forms (representational vehicle) with constant meaning (representational content).
In a series of experiments these differential predictions were put to a test. We briefly
summarise the results of the first three experiments and then report with proper
details the last, most encompassing study.
Stummer (1997, Experiment 2) developed the production version of the
“synonyms” task into two different versions of say-something-different (SSD)
tasks. In the categories task the alternative labels are a basic and a superordinate
category, i.e., if the puppet calls it “a dog” the child should call it “an animal” and
vice versa in response to the prompt “Now, say the other thing!” Stummer (2001,
Experiment 1) also used a subordinate and basic categories task (e.g., “poodle” vs.
“dog”). The “synonyms” and the categories tasks are both name/name (NN) tasks,
i.e., if puppet uses one name for the object, the child has to use the other name
for it. The colour/name (CN) task was designed as a control condition of similar
structure and complexity as the NN tasks but without the use of alternative names
(categories or synonyms). The child is instructed to say the object’s colour if the
puppet had said its name, and vice versa. For instance, if the puppet had said, “it’s
black,” then the child should say, “it’s a dog.” This task should be easier than the
other SSD tasks according to both theories. It should be easy according to ME
since the colour of an object does not specify what kind it is, hence there is no dual
CO
110
UN
109
J. Perner et al. / Cognitive Development 103 (2002) 1–22
OF
4
OF
J. Perner et al. / Cognitive Development 103 (2002) 1–22
5
Tasks
Stummer (1997)
(3.0–6.1, n = 36)
Stummer (2001)
(3.4–4.9, n = 40)
% Children correct on all items
False belief
61
Say-something-different
Synonyms
67
Categories
64
Colour/name
94
Part/name
–
Colour/colour
–
Part/part
–
73
–
73a
95
–
–
–
Present experiment
(2.8–4.9, n = 2 × 24)
Production
Judgement
19
19
38
42
92
96
87
92
33
21
87
92
87
92
.66∗∗ [.62∗∗ ]
.53∗∗ [.25]
.13 [−.10]
.11 [.27]
.17 [.19]
.15 [.18]
.72∗∗ [.72∗∗ ]
.60∗∗ [.32]
.17 [−.22]
.15 [.24]
.17 [−.08]
.15 [.32]
TE
Correlations [with age and verbal intelligence partialled out]
FB × synonyms
.64∗∗ [.45∗∗ ]
–
FB × categories
.59∗∗ [.42∗ ]
.77∗∗ [.65∗∗ ]
.37∗ [.20]
FB × colour/name
.38∗ [.32]
FB × part/name
–
–
FB × colour/colour
–
–
FB × part/part
–
–
DP
RO
Table 1
Results from experiments by Stummer and from the experiment reported here
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
RR
155
identity involved. It should be easy according to RUM since being a dog and being
black are two different states of affairs. The critical difference between theories is
that according to RUM the categories task should be as easy as the CN task while
according to ME it should be as difficult as the “synonyms” task.
The first 2 data columns in Table 1 present the results. The upper panel shows
the relative difficulty of the tasks in terms of the percentage of children who passed
each task, which clearly supports the expectations of the ME thesis: the categories
task and the synonyms task (both alternative naming NN tasks) were of equal
difficulty. The lower panel shows the correlations between tasks confirming the
results of Doherty (1994) and Doherty and Perner (1998) that the false belief task
relates strongly to the synonyms (and the categories task) even after differences in
age and verbal intelligence have been taken into account.
Stummer (2001, see Perner, Stummer, & Lang, 1999, Table 8.2) also found
that children had a general tendency to answer with the basic term (e.g., “dog”)
regardless of whether the correct alternative was a superordinate (animal) or a
subordinate category (poodle). This cannot be interpreted by RUM, which would
expect greater difficulty switching from the specific to the more general (e.g.,
poodle → dog) since there is no gain in meaning than from the general to the
more specific (dog → poodle) which can be more easily understood in terms of a
CO
154
UN
153
EC
a Average of number of children passing all superordinate/basic tasks and of number passing subordinate/basic tasks.
∗ P ≤ .05.
∗∗ P ≤ .01.
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
DP
RO
177
TE
176
EC
175
RR
174
difference in meaning. It does fit, however, the ME hypothesis that children prefer
a particular term for any one object.
Unfortunately for ME, Stummer’s data with the production paradigm also fit
another plausible explanation that the NN-tasks pose stronger executive demands
on children than the NC-task. It has been reported that mastery of the FB-task relates to improvements in inhibiting prepotent wrong responses (Russell, Mauthner,
Sharpe, & Tidswell, 1991) and several other executive tasks (Carlson & Moses,
2001; Hughes, Dunn, & White, 1998; for review see Perner & Lang, 1999). Hence,
a difference in executive demands provides an explanation for why the NN-tasks
are more difficult than the NC-task as well as for why the NN-tasks are mastered
at the same time as the FB-task. In particular, the NN tasks have a strong similarity
with the Stroop task, a prototypical executive function task. It is characterised by
the requirement to suppress the more automatised response (e.g., name the colour
designated by a colour word) to a request (name the colour of a colour word) in
order to respond with the desired alternative response (name the colour in which
the colour word is written). The NN-tasks pose the same executive demands. There
are two (or more) possible responses to the question, “What is this?” and the child
has to suppress the option that comes to mind first, when this option is the wrong
one. This is particularly strong when puppet has used the basic term and the child
has to use the super- or subordinate term (as observed by Stummer, 2001). It is also
a problem in the synonyms task because the name just highlighted by puppet (e.g.,
“rabbit”) is more likely to come to mind first and needs to be suppressed in order
to respond with the other option (bunny). In contrast, these executive demands are
not present in the NC-task where the child is instructed to switch between different
questions. For instance, if puppet answers the question about what the item is, then
the child has to answer the question about its colour (this task is comparable to
the Stroop control condition of naming the colour in which non-colour words are
written). We put this hypothesis to test by introducing a colour/colour (CC) and
a part/part task (PP). For instance, in the CC task puppet names one of the two
colours of the item and the child has to refrain from repeating the same colour and
name the other colour.
The executive demand hypothesis predicts that these tasks should be as difficult
as the NN-tasks because they create the same Stroop-like executive demands,
i.e., the same request, e.g., “name its colour!” or “name a part of it!” has to be
answered with an alternative possible answer, like in the NN-task. In contrast, the
ME hypothesis predicts that these new tasks should be as easy as the NC-task,
since colour or having a part are not alternative names for the object.
To strengthen the test we also included a judgement version as in the original
study by Doherty and Perner (1998). This also allows us to detect the possibility that executive demands determine children’s performance in conjunction with
representational understanding. In that case we expect the NN, CC and PP tasks to
be difficult in the production version because they all pose the presumed executive
demands, but only the synonyms task should pose difficulties in the judgement
version since the judgement version should not pose the Stroop type of execu-
CO
173
UN
172
J. Perner et al. / Cognitive Development 103 (2002) 1–22
OF
6
OF
J. Perner et al. / Cognitive Development 103 (2002) 1–22
7
219
3. Method
220
3.1. Participants
221
223
Forty-eight children (29 boys, 19 girls) from four nursery schools in Salzburg,
Austria participated in this study. Their ages ranged from 2.8 (2 years and 8 months)
to 4.9 (mean age of 3.11, S.D. = 7 months).
224
3.2. Design
225
240
Each child was tested in two sessions lasting roughly 15–20 min about 1 week
apart. Each session consisted of four tasks, one standard false belief (FB) task and
three SSD tasks. In one session the FB-task was administered before the SSD tasks
and in the other session it was administered after them. Also, in one session the
SSD tasks were administered in their production version, and in the other session
in their judgement version. These factors were fully counter-balanced. For half
of the children in each age group the three SSD tasks were a superordinate/basic
categories task (NN-categories), a name/colour task (NC) and a colour/colour task
(CC). For the other half the three tasks were a synonyms task (NN-synonyms), a
name/part task (NP) and a part/part task (PP). The order of the three SSD tasks was
fully counter-balanced. The orders of the three trial types of the judgement tasks
(puppet responds with same, different or wrong label) were also counter-balanced.
A child received a “pass” score for a particular condition if all three trial types
were answered correctly. For the production versions there were two tasks in
each condition (e.g., superordinate-basic, and basic-superordinate) and a child
was scored as passing if answers for both tasks were correct.
241
3.3. Materials
229
230
231
232
233
234
235
236
237
238
239
242
243
244
245
246
247
TE
228
EC
227
RR
226
The materials for the two false belief tasks were several Playmobile® figures.
In the dog story a dog puts his bone in one location and in his absence another
story character moves the bone unexpectedly to a new location. The book story
follows the same plot except that a boy Max puts his book in one location and in
his absence his mother moves the book unexpectedly to a new location. Moreover,
an extensive battery of memory questions was used at the end:
CO
222
Memory question 1:
Reality question:
Memory question 2:
Where did he put the bone in the beginning?
Where is the bone now?
Who put it there?
UN
217
DP
RO
218
tive demand. Whereas under the ME hypothesis, only the NN-tasks — but both
of them — should pose difficulties in the production as well as the judgement
version.
216
J. Perner et al. / Cognitive Development 103 (2002) 1–22
OF
8
251
3.4. Procedure
252
Each child was tested individually in a small room near the nursery school
classrooms. The procedures for the FB-task and the SSD-tasks were modelled
after the study by Doherty and Perner (1998). We only outline the procedure for
the judgement version.
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
3.4.1. Say-something-different (SSD) tasks — judgement version
For the vocabulary check 3 A4 sheets were used with four pictures on each. Two
of the pictures were experimental items used later (e.g., train/railway, stool/chair).
Children were asked to point to, e.g., a “train” (Zug), and then later to a “railway”
(Eisenbahn). The subsequent modelling phase differed for each kind of task. In the
NN tasks two labels were introduced, either two synonyms (“bunny” and “rabbit”)
or a basic and a superordinate category (rabbit and animal). Then the child was
asked what the item was called, and the child used one of the terms, e.g., “rabbit.”
It was then explained that the puppet, too, had to say what the item was called but
use a different name, i.e., “bunny.” The puppet shouldn’t say “rabbit,” because that
is what the child had called it. Puppet needs to use a different name, but one that
is right. Puppet couldn’t say “elephant” because this is not an elephant. Puppet
needs to say “bunny” because this is a bunny, and the child had not used “bunny.”
Then this game was practised on three trials. On the first trial puppet repeated the
same label the child had used, on the second trial the correct alternative was used
and on the third trial an incorrect term was applied (e.g., “elephant”). Each time
the child was asked: “Did Puppet say what he was supposed to say?1 ” Each time
the child gave a wrong answer, this was pointed out and the reason for why it was
wrong was repeated.
In the NC-task it was explained that the child can either name the item or say
what colour it has and that the puppet then had to say the other thing but not
something that was wrong. In the CC-task each object had two colours; children
could name one of them and puppet was supposed to name the other colour but
not one the object didn’t have. In the NP-task it was pointed out that the item was
a rabbit and that it had a head as a part. Then the child was supposed to either
name the item or the part and puppet had to provide the other label, but not one
that didn’t apply. In the PP-task two parts of the item were pointed out (head, tail),
the child had to name one of them and puppet the other one but not one the item
didn’t have. Each task was demonstrated in its three response versions, i.e., the
TE
256
EC
255
RR
254
CO
253
1 German original: “Hat Rex das gesagt, was er hätte sagen sollen?” NB: We avoided asking whether
it was right what puppet had said, because “right” might lead children to say “yes” when puppet repeated
the child’s correct label.
UN
249
DP
RO
250
The materials for the six SSD-tasks consisted of several A4 sheets with a single
object depicted on it. A list of all items used for the test-phases in the production
and judgement version of the SSD-tasks is given in the Appendix A.
248
OF
J. Perner et al. / Cognitive Development 103 (2002) 1–22
9
puppet wrongly repeating what the child had said, correctly mentioning the other
option, or incorrectly applying a wrong label.
287
3.5. Results
288
3.5.1. False belief
Children’s answers to the reality control question (Where is the object now?)
were almost perfect (90% were correct in both tasks). They were a little less
reliable in their answers to the two memory control questions: only 65% gave
correct answers in both tasks. Nevertheless, 30 of the 48 children (62%) answered all three control questions in both stories correctly, but only 9 of the
48 children (19%) answered both test questions correctly. Since only two children answered one of the test questions correctly and the other incorrectly, the
retest reliability (κ = .874, P < .001) of test question answers was very
good.
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
TE
293
3.5.2. Say-something-different
Children were almost perfect in the vocabulary check (96% correct). The percentage of children who passed both tasks of the test phase is shown in Table 1
(upper panel fourth and fifth column). Children’s problems are restricted to the
two NN-tasks with either synonyms or super-basic categories. There was no difference between the group with synonyms and the group with categories, neither
in the production version: χ 2 (1, N = 48) = .09, P > .76, nor in the judgement version: χ 2 (1, N = 48) = .95, P > .33. Although the higher percentages
in column 4 than in column 5 indicate, in particular for the synonyms task, that
more children passed the production task than the judgement version; the difference is, however, not statistically significant (binomial test: n = 8, k = 1,
P > .05).
The remaining percentages in Table 1 show that children did exceedingly well on
all the other SSD-tasks (NC, NP, CC, and PP). There was no significant difference
in difficulty between them. There was no difference between the group being tested
on colour and the group being tested with parts: ␹2 (1, N = 48) = .36, P > .55,
for NC versus NP in the production version, and χ 2 (1, N = 48) = .22, P > .63,
in the judgement version, and χ 2 (1, N = 48) = .22, P > .63, for CC versus PP
in the production version, and χ 2 (1, N = 48) = .22, P > .63, in the judgement
version. Also, children did no better on the tasks with names (NC or NP) than
on tasks without names (CC or PP): McNemar χ 2 (1, N = 48) = 1.0, P > .62,
for the production version and McNemar χ 2 (1, N = 48) = .00, P > .99, in the
judgement version.
Fig. 1 shows the developmental trend on these groups of SSD-tasks and the
FB-tasks by splitting the group of children in three age groups. It makes clear that
even the youngest children performed close to ceiling on all SSD-tasks except on
the ones that involved dual naming (NN-tasks). On these tasks a strong developmental improvement was apparent that paralleled the usual improvement found on
EC
292
RR
291
CO
290
UN
289
DP
RO
286
285
OF
J. Perner et al. / Cognitive Development 103 (2002) 1–22
TE
DP
RO
10
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
RR
328
the FB-task in this age range (Wellman, Cross, & Watson, 2001). There was significant improvement with age on all three tasks: χ 2 (2, N = 48) = 9.76, P < .005,
for NN-production, χ 2 (2, N = 48) = 6.85, P < .01, for NN-judgement, and
χ 2 (2, N = 48) = 11, P < .01, for FB.
The correlation coefficients in the lower panel of Table 1 (columns 4 and 5) show
the NN-tasks and the FB-task also correlate very strongly with each other. These
correlations remain robust for synonyms even when age and verbal intelligence
(K-ABC) are partialled out (numbers in square brackets).
The essence of the results stays the same when the analysis is reduced to those
children who answered all control questions correctly. Of these children 20%
passed the FB-task, 23% the NN-judgement and 37% the NN-production task and
97% all of the other SSD-tasks in the judgement version and 100% all of the other
SSD-tasks in the production version. The difference between NN-tasks and other
SSD-tasks was significant for both production (McNemar’s χ 2 (1, N = 30) = 19,
P < .001) and judgement versions (McNemar’s χ 2 (1, N = 30) = 22, P <
.001). The correlation between FB- and the NN-judgement task is then r = .71,
P < .001, [pr = .67, P < .001, with age and verbal intelligence partialled out],
and between FB and the NN-production task r = .66, P < .001, [pr = .58,
P < .005].
CO
327
UN
326
EC
Fig. 1. Percentage of children in the three age groups passing each experimental condition.
OF
J. Perner et al. / Cognitive Development 103 (2002) 1–22
11
3.6. Discussion
346
371
The results are exactly as predicted from the ME hypothesis: both NN-tasks,
synonyms as well as categories (both involve alternative labels for an object), are
as difficult as the false belief task in their production as well as the judgement
version. All other SSD-tasks are equally easy.
The massive difference between children’s performance on the NN-tasks and
the other SSD-tasks speaks clearly against the executive demand hypothesis. It also
makes our results difficult to explain by other potential alternatives, for instance,
the CCC theory (cognitive complexity and control theory, Zelazo & Frye, 1997).
The core claim is that children around 4 years old come to master embedded
conditionals, and the false belief task has this logical structure (Frye, Zelazo, &
Palfai, 1995; though see Perner, Stummer, & Lang, 1999, for reservations about
how this analysis applies to the FB-task). The same analysis could be brought to
bear on the SSD tasks (Perner et al., 1999). This could explain why the NN-tasks
are mastered at the same time as the FB-task. However, it is difficult to see how
it could account for the fact that the other SSD tasks are substantially easier,
since their logical structure seems identical. Similar problems arise for so-called
“neo-Piagetian” theories (Case, 1985; Pascual-Leone, 1970) or other theories that
argue with developmental changes in working-memory limitations, because it is
difficult to see how the different SSD tasks could pose different working-memory
loads.
In sum, the data from this experiment confirm the problems for RUM theory,
and they also speak against the executive demands hypothesis and other potential
alternatives. This leaves the ME hypothesis as the only viable alternative for explaining children’s difficulty with the NN-tasks. However, we need to look again
at the reasons, why the ME constraint — the reluctance to accept more than one
label for an object — should be related to the inability to understand belief.
372
4. Problems with Flavell’s and Markman’s “single identity” hypothesis
373
Although the data on the SSD-tasks clearly speak for the ME hypothesis, the
question remains why ME should be developmentally related to failure of understanding visual perspective, the appearance–reality distinction and — in our
case — false belief. Markman (1989) suggested that ME operates not at the level
of linguistic labels but results from deeper principles of categorisation, namely
that “children may believe that an object has one and only one identity — that it
can be only one kind of thing . . . .” (Markman, 1989, p. 212). Flavell suggested
that the “single identity hypothesis” might also explain children’s problems with
appearance–reality problems where, e.g., a piece of sponge appears to be a rock.
“Children of this age also believe . . . that each object or event in the world has
only one nature — one ‘way that it is’ — at any given point in time. It cannot be
two or more very different, mutually contradictory, and incompatible things at the
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
374
375
376
377
378
379
380
381
382
383
384
TE
351
EC
350
RR
349
CO
348
UN
347
DP
RO
345
J. Perner et al. / Cognitive Development 103 (2002) 1–22
OF
12
411
5. What is a perspective?
412
The Concise Oxford English Dictionary (Thompson, 1995) gives three relevant
definitions:
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
413
414
415
416
417
418
419
420
421
422
423
424
TE
390
EC
389
RR
388
(1) the art of drawing solid objects . . . so as to give the right impression of relative
positions, size, etc.; (2) the apparent relation between visible objects as to position,
distance, etc.; (3) a mental view of the relative importance of things (keep the right
perspective).
CO
387
Generalising these definitions in representational terms we might say that a difference in perspective occurs whenever there is a difference in content between
different representations of a particular thing (an object, a scene or an event).
Three aspects are critical for a difference in perspective. (1) There must be
some difference and incompatibility (e.g., different apparent relations). (2) The
difference has to do with how the target is represented (representational reference).
(3) Despite their difference the representations must pertain, in some sense, to the
UN
386
DP
RO
410
same time; rather, it can only be one thing. Consequently, it makes no sense to
them to hear something described as being radically different than the single way
it ‘is’ (with ‘is’ not differentiated from ‘seems to them at that moment’)” (Flavell,
1988, p. 245).
It remains far from clear how the “single identity” hypothesis is to be applied
to the false belief task. It raises the question about which object or event has only
one nature. One might suggest the object’s location as a candidate event. Indeed,
this event is given two different labels, e.g., “drawer” (in the false belief) and
“cupboard” (in reality). However, there are several problems with this analysis. One
is children’s surprisingly early competence in pretence (e.g., Harris & Kavanaugh,
1993) where children as young as 2.5 years have little problems pretending that
a block (in reality) is a soap bar (in pretence). So why can they understand dual
identity (block–soap) in this case but not in the belief task?
Another problem regards Flavell’s characterisation of what children fail to understand: “It cannot be two or more very different, mutually contradictory, and
incompatible things at the same time.” This doesn’t apply as an explanation to
the false belief task (or the appearance–reality distinction), because children are
not required to understand that the object’s location can be the drawer and the
cupboard at the same time. And it doesn’t apply to the NN-tasks because being
a rabbit and being an animal are not mutually contradictory. Rather, the question
to be answered about ME is not how children manage to understand that a thing
(object or event) can be two incompatible things but why children treat compatible
(e.g., rabbit, animal) as incompatible. To answer this question we look at children’s
understanding of perspective as a common denominator for their problems with
false belief and their tendency to show ME. Before we can achieve this we need
to get clear on what we mean by “perspective.”
385
OF
J. Perner et al. / Cognitive Development 103 (2002) 1–22
13
460
5.1. Multiple identity
461
By using different words to say what something is, we do not primarily predicate different properties to something but we apply alternative sortals (words that
determine what sort of thing something is) and thereby individuate the very thing
we are talking about in different ways (e.g., Hirsch, 1982). A classical example
that brings out the problems created by the use of different sortals is the case of
the statue made of clay (e.g., Perry, 1970). Let’s say you gave me a piece of clay
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
462
463
464
465
466
TE
430
EC
429
RR
428
CO
427
UN
426
DP
RO
459
same target (one and the same ‘thing’ that is represented). We only face a difference
in visual perspective if we are looking at the same group of objects from different
vantage points, not if we were looking at totally different objects. Or the word
chosen for it only gives a different perspective from another word chosen for it,
if the “it” refers to the same ‘thing’ in both cases (conceptual perspective taking,
Clark, 1997; Tomasello, 1999).
It seems clear how this applies to visual perspective differences and to false
beliefs. If we face each other and between us on a line are a tree and a rock
then I see the tree in front of the rock while you see it behind the rock. Our two
corresponding statements “tree in front of rock,” and “tree behind the rock” are
incompatible in the sense that one of us could not sincerely utter both of them. Our
two statements are apparently about the same target, i.e., the same two objects and
their relative spatial relation between them. Finally, to make sense of what is going
on we need to refer to our different points of view (representational reference).
Similarly for false belief, our knowledge of where the chocolate really is (chocolate
in location B) and Max’s claim that it is still in A are clearly incompatible, they
are about the same target (the real location of the chocolate), and to make sense
of them we have to identify Max’s claim as one of a false belief (representational
reference).
Now for an example that we would not call a perspective difference. We are
looking at a Dalmation. I say, “It is beautifully spotted” and you say, “It is a dog.”
Does this involve a perspective difference? It involves alternative representations
(descriptions) but there is no incompatibility, as I can immediately take your statement on board and say, “It is a dog and it is beautifully spotted.” It also brings out
a problem in deciding sameness of target. If we consider the dog as the target, then
we are speaking about the same target, but if we define the target in a more fine
grained way then you are speaking about its nature, and I about its colouring.
But now to the critical issue: What if you call it a dog and I a Dalmatian? Several
authors assume that this creates a difference in perspective: in each conversation,
it is said, we develop a particular conceptual perspective on the world, one of
the central aspects of language acquisition (Tomasello, 1999). The most typical
examples involve “multiple identity,” e.g., describing a dog as a Dalmatian, dog, or
animal (Clark, 1997). Does such alternative naming create different perspectives
according to our criteria? To see that it does we need to go more deeply into what
we do when we label objects.
425
J. Perner et al. / Cognitive Development 103 (2002) 1–22
OF
14
488
6. Two types of perspective and the “same thing” problem
489
6.1. Truth-incompatible perspectives
490
501
False beliefs (the chocolate is in location A) vis-à-vis reality (the chocolate is in
B) are clear cases of a difference between truth-incompatible perspectives. That’s
why these beliefs are called false. Similarly, that is why deceptive appearances,
e.g., the piece of sponge appears to be a rock, are called deceptive, because the
appearance misspecifies reality. In these cases it is clear that, e.g., Max’s mistaken
statement, “the chocolate is in A,” and the child’s knowledge, “the chocolate is
in B,” have the same target, namely the actual location of the chocolate, although
Max’s belief misrepresents that target. In case of doubt about what the target
precisely is, one can define it as that against which the truth of the propositions,
“the chocolate is in A,” and, “the chocolate is in B,” have to be evaluated, i.e.,
where the chocolate really is. No such clear criterion for determining the target is
available for truth-compatible perspectives.
502
6.2. Truth-compatible perspectives
503
Visual perspectives and conceptual perspectives are prime examples of this
class. They share the feature that alternative perspectives are true. For instance,
473
474
475
476
477
478
479
480
481
482
483
484
485
486
491
492
493
494
495
496
497
498
499
500
504
TE
472
EC
471
RR
470
CO
469
UN
468
DP
RO
487
last week, which an artist friend of mine turned into a lovely statue yesterday. Now
when faced with the statue I can individuate it as a piece of clay (the piece I got
from you last week) or as a statue (the one made yesterday by my friend). On first
blush and by normal intuition it seems that we have a single object that is a piece of
clay as well as a statue — just like it is lengthy and brownish. However, this is not so
straightforward. When asked since when I have been in possession of it, the answer
depends on what sortal I used to individuate it. If I used “piece of clay” the answer
is “last week,” if I used “statue” the answer is “yesterday.” So, here we have our incompatibility resulting from alternative individuation by different labels (sortals).
One attempt at explaining what is going on is to assume that there are actually
two different things in existence, a piece of clay and a statue that temporally
overlap, i.e., a part of the temporal ‘stage’ of the piece of clay and the ‘stage’ of
existence of the statue are identical (e.g., Perry, 1970). In that case, there would not
be a difference in perspective involved, since the seemingly incompatible answers
about ‘its’ existence are not incompatible after all, because “it” refers to different
things. However, the question remains, is that just a philosopher’s way out of a
quandary or do we actually think that way (despite our conscious denial of it)?
We see, the question whether two representations have the same target or different targets and also the question whether their contents are incompatible or
not, are sometimes difficult to answer. To see where the difficulty lies we need to
distinguish two different types of perspective difference.
467
OF
J. Perner et al. / Cognitive Development 103 (2002) 1–22
15
524
7. A constructivist definition of perspective
525
At the starting point we use a technical trick and assume all representations
are representations of the same ‘thing:’ the logical universe (the set of all possible worlds). That means that any two representations with different content are
different perspectives on the universe. For instance, both of us look at the same
Dalmatian and one of us forms the thought, “it is a dog,” and the other, “it is
perfectly spotted.” But there is no need, yet, to speak of different perspectives,
if we can give the universe more structure (construct parts). That this is easily
possible shows in the fact that our two thoughts can be easily integrated within
a single thought, “it is a dog AND it is perfectly spotted,” or even, “it is a perfectly spotted dog.” In other cases this integration requires more construction. The
two statements, “It is cold and raining” (spoken in Salzburg) and, “it is hot and
sunny,” (spoken in Philadelphia) seem incompatible until we realise that they refer
to different parts of the world. One describes the weather in Salzburg, the other
the weather in Philadelphia. Consequently, they can be uttered by a single person
making the difference in target explicit: “It is cold and raining in Salzburg and it
is hot and sunny in Philadelphia.”
Similar cases with a hidden difference in target are often found when people
refer to different times or different worlds. If you say, “Santa Claus doesn’t exist”
and I, “Santa Claus lives near the North Pole,” we do not need to construe this as
different perspectives on a single universe, but we can understand it as talk about
different worlds, the real world in which Santa Claus does not exist and a fictional
511
512
513
514
515
516
517
518
519
520
521
522
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
TE
510
EC
509
RR
508
CO
507
UN
506
DP
RO
523
spatial descriptions given from different vantage points like, “A is in front of B,”
and “A is behind B,” cannot both be uttered by the same person, “A is in front
of B and A is behind B,” hence there is some incompatibility, but if uttered from
different vantage points neither speaker can be excused of being wrong, i.e., both
statements are true. Similarly in the case of applying alternative sortals to one and
the same entity: if the statements, “This is a statue,” and “this is a piece of clay,”
create different perspectives, as some would argue (e.g., Clark, 1997; Tomasello,
1999), then they are truth-compatible perspectives, since statements conforming to
either perspective, e.g., “I have it since yesterday,” and “I have it since last week,”
are both true.
The problem with a single target is now the following. Although at face value,
“A is in front of B” and “A is behind B” are incompatible, they cannot be incompatible in some absolute sense, or else they could not both be true. Two solutions
to this puzzle are possible. Either there is a hidden difference in target, and they
are only superficially incompatible, or they have the same target (the spatial relation between A and B) but their descriptors are applicable only within different
frames of reference (Kutschera, 1999, suggested that solution for the dual identity
problem) and are incompatible within a particular frame of reference. We adopt a
constructivist approach to deciding which option to take.
505
J. Perner et al. / Cognitive Development 103 (2002) 1–22
OF
16
563
8. Children’s mastery of perspective taking
564
We need to make an important distinction between different levels of “perspective taking:” switching perspectives (taking different perspectives at different times) and confronting perspectives (representing two perspectives simultaneously; understanding that there are different perspectives). In line with Clark’s
(1997) many-perspectives view, which she contrasts with the one-perspective view
ascribed to defenders of ME, children can switch conceptual perspectives more or
less from the time they start building up a larger vocabulary in the second year of
life. Once children know that a rabbit is an animal, they can look at it as a rabbit
and switch to considering it an animal. But they need not confront the fact that it
can be considered an animal as well as a rabbit.
Our claim (compatible with Clark’s) is that the ability to confront different
perspectives emerges around 4 years and underlies the co-emergence of success
on the FB and the NN-tasks (and, of course, other perspective tasks: Level 2
visual perspective taking and appearance–reality, Flavell, 1988; Perner, 1991). This
locates the acquisition a few years earlier than Piaget and Inhelder (1948/1956).
However, if one strips their methods of unnecessary complexities (Light & Nix,
1983) then the age of acquisition comes closer to the age that we are dealing with
here. The characteristic feature of the NN-tasks is that one needs to keep both
descriptions in mind in order to judge whether the other person (puppet) has used
the alternative description. If alternative descriptions create different perspectives
then children must be able to confront different perspectives in order to make
these judgements. The same ability is needed in the false belief task in order to
represent that the object is in location 2, but is believed to be in location 1. These
552
553
554
555
556
557
558
559
560
561
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
TE
551
EC
550
RR
549
CO
548
UN
547
DP
RO
562
world in which he is known to live near the North Pole. Evidently, our constructive powers to integrate different mental contents as stemming from a coherent
logical universe with different parts (different worlds) are considerable. Adopting Piaget and Inhelder’s (1941/1974, p. 87) language for describing intellectual
development: we progress beyond a flow of independent perspectives (egocentric phenomenalism) by discovering properties of the logical universe (world) as
objective co-ordinations among these perspectives.
But there comes a point where we do not have the conceptual structures to find
such a solution for integrating seemingly incompatible contents. Spatial perspectives are the classic case. “A is in front of B” and “A is behind B” cannot be simply
integrated by making a hidden difference in target explicit, e.g., “A is in front of
B
and A is behind B
” The only common sense way of filling in the
missing parts is with a representational reference to different frames of reference
or points of view, e.g., from my/your point of view, from where I am/you are, etc.
That is, our only natural way of understanding the relationship between the two
statements is to interpret them as pertaining to different perspectives (points of
view, frames of reference).
546
OF
J. Perner et al. / Cognitive Development 103 (2002) 1–22
17
two propositions can be integrated in a single representation only by marking them
as different perspectives.
589
9. Mutual exclusivity — a problem of dual identity or perspective?
590
It is now tempting to conclude that the ability to represent differences in perspective frees children from the ME bias. This is not necessarily the case. For,
from realising that I can conceive of a ‘thing’ under different sortals within different perspectives it does not follow that the ‘thing’ can have both identities. If
this did follow, we would have to conclude from the fact that an object in the
cupboard can be (mistakenly) conceived of as being in the drawer, that it must
be possible that an object can be in the cupboard and in the drawer at the same
time.
Clearly, we need to take into account the difference between truth-compatible
and truth-incompatible perspectives. The NN-task requires the ability to represent
two perspectives simultaneously, but does not require them to evaluate the truth of
both claims. In a new rejection task Doherty (2001) asked children between 3 and
8 years to do exactly that. Children were first asked what something (e.g., a rabbit)
is called. If they answered “rabbit” then a doll named Sally also called it a “rabbit”
but Tony called it a “bunny.” Children then had to remember what Tony had said
(bunny) and they were asked: “Is it a bunny?” The same procedure was repeated
for what Sally had said it was. Children had few problems remembering what each
character had said, but many children rejected one of the two labels (they didn’t
think it could be a bunny and a rabbit) which is indicative of the classical ME
effect.
Not all children who failed the NN-task tended to reject (as one might have
thought). This can be expected if some children were induced to switch perspectives with each switch to the other sortal. More interestingly, rejection didn’t stop
altogether with success on the NN-task around 4 or 5 years. Only around 7.5
years did the rejection rate get close to zero. This considerable lag between success on the NN-task and subsidence of ME can be explained by fact that the
NN-task requires the ability to confront a difference in perspective while ME can
only be overcome if the truth-compatibility of two perspectives can be assessed in
addition.
The data, therefore, suggest that truth-compatibility is not understood before
the age of 7–8 years. This gives us two interesting testable predictions. The first
prediction is that a similar lag should be observed for other truth-compatible perspectives, e.g., visual perspective. Extrapolating from existing research (Kuczaj &
Maratsos, 1975; Tanz, 1980) we assume that by 4 years of age children might
be able to understand that when A is in front of B for them, that another person viewing the scene from the opposite direction will say, “A is behind B.”
However, when asked, “Is that right what other said?” children until 8 years will
reject it.
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
TE
595
EC
594
RR
593
CO
592
UN
591
DP
RO
588
587
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
DP
RO
633
TE
632
EC
631
RR
630
Another interesting prediction can be derived from the consideration that understanding truth-compatibility requires a kind of higher-order, perspective-relative
truth evaluation. The easiest way of deciding whether another person, who is
claiming something different, may be right or wrong is to simply judge the truth
of other’s statement from within one’s own perspective. This works nicely for the
false belief task. When Max on his return exclaims, “the chocolate is in the drawer,”
a check within my own perspective, where the chocolate is in the cupboard, leads
to the correct conclusion that his statement is false. The same strategy, however,
does not work for visual perspectives. If we face each other over objects A and
B, and I evaluate your statement, “A is in front of B,” within my perspective then
I come to the wrong conclusion that your statement is false, because for me B is
in front of A. And, it does not work for conceptual perspective changes arising
from alternative naming, because when one is taking the perspective where it is
being individuated as a statue then it is not a piece of clay (with ‘is’ in the sense
of individuation).
A higher-order strategy of evaluating truth within your perspective is needed
here. Data from a belief task used by Perner and Howes (1992) and Perner (1994)
point to a later age of acquisition. Children had to predict what Max would answer to
the question, “Where is your chocolate?” As one would expect from previous data,
practically all 4–6-year olds predicted correctly his mistake. However, children
also were asked to predict what Max would say to the question, “Do you know
where your chocolate is?” Now many of these children responded wrongly that
he would answer “No, I don’t know.” The main explanation for the difficulty of
this question is that the answer requires the second-order understanding that Max
will reflect (from within his perspective) approvingly on the truth of his belief,
in other words, he thinks he knows. On the basis of these results we can venture
the following predictions. Children will misjudge the truth of truth-compatible
statements as false until they are able to master the second-order strategy at about
6 years. Until that age, many children are expected to misjudge another person’s
true spatial statement phrased within an egocentric reference frame (e.g., “A is in
front of B”) when spoken from a different vantage point as false, and will reject
the application of one of two alternative sortals in Doherty’s rejection task.
This, however, cannot be the end of the developmental story. The child must
also learn when the truth needs to be evaluated within her own perspective and
when it needs to be evaluated within the other’s perspective. If every statement
is evaluated within the other’s perspective then false beliefs will be misjudged
as true. Plausibly this distinction is acquired with the ability to evaluate truth
relative to other perspectives. However, the possibility exists that this acquisition
may take additional years leading to a period of children’s regression in their
ability to judge the falseness of false beliefs, and a stage of uncertainty about
rejecting synonyms. Such uncertainty could be the reason for Doherty’s finding
that some children keep occasionally rejecting synonyms up to the age of 7 or even 8
years. But these speculations are in need of systematic and, preferably, longitudinal
investigation.
CO
629
UN
628
J. Perner et al. / Cognitive Development 103 (2002) 1–22
OF
18
OF
J. Perner et al. / Cognitive Development 103 (2002) 1–22
19
10. A new perspective including class inclusion
673
10.1. Back to Piaget
674
690
The ability required in the alternative naming tasks to acknowledge that something can be a rabbit and an animal at the same time should be a critical prerequisite for passing Inhelder and Piaget’s (1964) classic class inclusion test. For, asked
about a group of three rabbits and two cats, whether there are more rabbits or more
animals, one has to count the rabbits as rabbits and as animals. Our analysis claims
that two perspectives are involved. Independent counts (How many rabbits? How
many animals?) can be provided by switching perspectives. Children (provided
they can count) should be able to do this before they master alternative naming
and the false belief task. However, the class inclusion question asks for comparison
of both sets and so both perspectives need to be confronted. Moreover, children
need to be aware that the numbers they attain within each perspective are both
accurate (true), which suggests that children solve the class inclusion problem at
the same time as they can answer the know-question by Perner and Howes (1992)
and when they stop rejecting on Doherty’s rejection task. To settle this issue we
would need more information of how children approach the class inclusion task.
The variability found with different task procedures (Winer, 1980) might be due
to how these procedures influence children’s approach to the task.
691
11. A Piagetian epilogue
692
700
In our attempt to explain why alternative naming and understanding false belief are developmentally related we made use of several Piagetian ideas. (1) We
used Piaget’s constructivist view of intellectual development for our definition of
perspective. (2) We used Piaget’s and Inhelder’s notion that an understanding of
perspective is acquired in early-to-middle childhood as a powerful tool to integrate different developments at the age of 4 years. (3) By differentiating between
truth-compatible and truth-incompatible perspectives we were able to give a new
— and hopefully productive — perspective on mutual exclusivity and how it relates
to Inhelder and Piaget’s class inclusion task.
701
Acknowledgments
702
We would like to thank the staff and children of the Magistrat Kindergärten
in Auwiesenstrasse 22-24 and 60, Neutorstrasse, Maxglan, the Pfarrkindergärten
Herrnau, Nonntal, Parsch and the Gemeindekindergärten Radstadt and Ramsau
for their participation in these studies. The project was financially supported by
the Austrian Science Fund (Project P14495-SPR) and by a research grant from the
Faculty of Natural Sciences, University of Salzburg, to Manuel Sprung.
680
681
682
683
684
685
686
687
688
689
693
694
695
696
697
698
699
703
704
705
706
707
TE
679
EC
678
RR
677
CO
676
UN
675
DP
RO
672
J. Perner et al. / Cognitive Development 103 (2002) 1–22
Appendix A
709
Items for the test phase of the say-something-different tasks.
DP
RO
708
OF
20
Production version
Judgement version
Synonyms
Woman/lady [Frau/Dame]
Aircraft/plane
[Flugzeug/Flieger]
Train/railway [Zug/Eisenbahn]
Matches/machsticks
[Zünd-/Streichholz]
Stool/chair [Sessel/Stuhl]
Categories
Pear/fruit [Birne/Obst]
Rabbit/animal [Hase/Tier]
Pliers/tool [Zange/Werkzeug]
Apple/fruit [Apfel/Obst]
Cat/animal [Katze/Tier]
Colour/name
Blue/cup [Blau/Tasse]
Green/snail [Grün/Schnecke]
Red/car [Rot/Auto]
Blue/flower [Blau/Blume]
Red/ball [Rot/Ball]
Part/name
Donkey/ears [Esel/Ohren]
Snowman/nose
[Schneemann/Nase]
TE
Tasks
Elephant/snout [Elefant/Rüssel]
Monkey/tail [Affe/Schwanz]
Dragon/head [Drache/Kopf]
710
Yellow/blue bee [Gelb/Blau]
EC
Part/part
Yellow/green tree
[Gelb/Grün]
Yellow/blue fish [Gelb/Blau]
Head/tail [Kopf/Schwanz]
Wing/head [Flügel/Kopf]
Green/red present [Grün/Rot]
Blue/red tricycle [Blau/Rot]
Nose/mouth [Nase/Mund]
Head/tail [Kopf/Schwanz]
Head/belly [Kopf/Bauch]
RR
Colour/colour
714
References
715
Carey, S. (1985). Conceptual change in childhood. Cambridge, MA: MIT Press (A Bradofrd Book).
Carey, S., & Bartlett, E. J. (1978). Acquiring a single new word. Papers and Reports on Child Language
Development, 15, 17–29.
Carlson, S. M., & Moses, L. J. (2001). Individual differences in inhibitory control and children’s theory
of mind. Child Development, 72, 1032–1053.
Case, R. (1985). Intellectual development: Birth to adulthood. Orlando, FL: Academic Press.
Clark, E. V. (1997). Conceptual perspective and lexical choice in acquisition. Cognition, 64, 1–37.
716
717
718
719
720
721
UN
712
CO
713
Note: The German originals are in the square brackets. The pairs used in the
synonyms task are used fairly synonymously in German, which is not the case for
their English translations.
711
21
CO
RR
EC
TE
DP
RO
Clements, W.A., Heider, A., Brooks, J., & Garnham, A. (1998). From synonyms to homonyms:
Exploring the role of metarepresentation in language understanding. Unpublished manuscript,
University of Sussex.
Dockrell, J., & Campbell, R. (1986). Lexical acquisition strategies in the preschool child. In S. A.
Kuczaj & M. D. Barrett (Eds.), The development of word meaning. New York: Springer-Verlag.
Doherty, M. (1994). Metalinguistic understanding and theory of mind. Unpublished doctoral
dissertation, Laboratory of Experimental Psychology, University of Sussex.
Doherty, M. (2001). Rejection of synonyms: A novel mutual exclusivity phenomenon. Unpublished
manuscript, University of Stirling, Scotland.
Doherty, M., & Perner, J. (1998). Metalinguistic awareness and theory of mind: Just two words for the
same thing? Cognitive Development, 13, 279–305.
Flavell, J. H. (1988). The development of children’s knowledge about the mind: From cognitive
connections to mental representations. In J. W. Astington, P. L. Harris, & D. R. Olson (Eds.),
Developing theories of mind (pp. 244–267). Cambridge: Cambridge University Press.
Flavell, J. H., Flavell, E. R., & Green, F. L. (1983). Development of the appearance–reality distinction.
Cognitive Psychology, 15, 95–120.
Forguson, L., & Gopnik, A. (1988). The ontogeny of common sense. In J. W. Astington, P. L. Harris, &
D. R. Olson (Eds.), Developing theories of mind (pp. 226–243). Cambridge: Cambridge University
Press.
Frye, D., Zelazo, P. D., & Palfai, T. (1995). Theory of mind and rule-based reasoning. Cognitive
Development, 10, 483–527.
Garnham, W. A., Brooks, J., Garnham, A., & Ostenfeld, A.-M. (2000). From synonyms to homonyms:
Exploring the role of metarepresentation in language understanding. Developmental Science, 3,
428–441.
Harris, P. L., & Kavanaugh, R. D. (1993). Young children’s understanding of pretence. Society for
Research in Child Development Monographs (Serial No. 237).
Hirsch, E. (1982). The concept of identity. New York, Oxford: Oxford University Press.
Hughes, C., Dunn, J., & White, A. (1998). Trick or treat? Uneven understanding of mind and emotion
and executive dysfunction in ‘hard to manage’ preschoolers. Journal of Child Psychology and
Psychiatry, 39, 981–994.
Inhelder, B., & Piaget, J. (1964). The early growth of logic in the child: Classification and seriation.
New York: Norton.
Kuczaj, S. A., & Maratsos, M. (1975). On the acquisition of front, back and side. Cognitive
Development, 46, 202–210.
Kutschera, F. V. (1999). Probleme der Identität. Facta Philosophica, 1, 3–19.
Leslie, A. M. (1994). Pretending and believing: Issues in the theory of ToMM. Cognition, 50, 211–238.
Light, P., & Nix, C. (1983). “Own Vew” versus “Good View” in a perspective-taking task. Child
Development, 54, 480–483.
Markman, E. M. (1989). Categorization and naming in children: Problems of induction. Cambridge,
MA: A Bradford Book.
Markman, E. M., & Wachtel, G. F. (1988). Children’s use of mutual exclusivity to constrain the
meanings of words. Cognitive Psychology, 20, 121–157.
Merriman, W. E., & Bowman, L. L. (1989). The mutual exclusivity bias in children’s word learning.
Monographs of the Society for Research in Child Development, 54 (3–4, Serial No. 220).
Pascual-Leone, J. (1970). A mathematical model for the transition rule in Piaget’s developmental
stages. Acta Psychologica, 32, 301–345.
Perner, J. (1988). Developing semantics for theories of mind: From propositional attitudes to mental
representation. In J. W. Astington, P. L. Harris, & D. R. Olson (Eds.), Developing theories of mind
(pp. 141–172). New York: Cambridge University Press.
Perner, J. (1991). Understanding the representational mind. Cambridge, MA: MIT Press, A Bradford
Book.
Perner, J. (1994). The necessity and impossibility of simulation. Proceedings of the British Academy,
83, 129–144.
UN
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
OF
J. Perner et al. / Cognitive Development 103 (2002) 1–22
CO
RR
EC
TE
DP
RO
Perner, J., & Howes, D. (1992). “He thinks he knows” and more developmental evidence against the
simulation (role-taking) theory. Mind & Language, 7, 72–86.
Perner, J., & Lang, B. (1999). Development of theory of mind and executive control. Trends in Cognitive
Sciences, 3, 337–344.
Perner, J., Stummer, S., & Lang, B. (1999). Executive functions and theory of mind: Cognitive
complexity or functional dependence? In P. D. Zelazo, J. W. Astington, & D. R. Olson (Eds.),
Developing theories of intention: Social understanding and self-control (pp. 133–152). Mahwah,
NJ: Lawrence Erlbaum.
Perry, J. (1970). The same F. Philosophical Review, 79, 181–200.
Piaget, J., & Inhelder, B. (1948). The child’s conception of space. London: Routledge and Kegan Paul.
Piaget, J., & Inhelder, B. (1974). The child’s construction of quantities: Conservation and atomism.
London: Routledge & Kegan Paul.
Russell, J., Mauthner, N., Sharpe, S., & Tidswell, T. (1991). The ‘windows task’ as a measure of strategic
deception in preschoolers and autistic subjects. British Journal of Developmental Psychology, 9,
331–349.
Stummer, S. (1997). Metalinguistisches Bewubtsein und “Theory of Mind.”
Stummer, S. (2001). Sag es anders — ToM! Von einem metalinguisitschem Bewusstsein zu alternativen
Erklärungsansätzen.
Tanz, C. (1980). Studies in the acquisition of deicitc terms. Cambridge: Cambridge University Press.
Thompson, D. (Ed.). (1995). The concise Oxford dictionary of current English. Oxford: Clarendon
Press.
Tomasello, M. (1999). The cultural origins of human cognition. Cambridge: Harvard University Press.
Wellman, H. M., Cross, D., & Watson, J. (2001). Meta-analysis of theory of mind development: The
truth about false belief. Child Development, 72, 655–684.
Wimmer, H., & Perner, J. (1983). Beliefs about beliefs: Representation and constraining function of
wrong beliefs in young children’s understanding of deception. Cognition, 13, 103–128.
Winer, G. A. (1980). Class-inclusion reasoning in children: A review of the empirical literature. Child
Development, 51, 309–328.
Zaitchik, D. (1990). When representations conflict with reality: The preschooler’s problem with false
beliefs and “false” photographs. Cognition, 35, 41–68.
Zelazo, P. D., & Frye, D. (1997). Cognitive complexity and control: A theory of the development of
deliberate reasoning and intentional action. In M. Stamenov (Ed.), Language structure, discourse
and access to consciousness (pp. 113–153). Amsterdam and Philadelphia: John Benjamins.
UN
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
J. Perner et al. / Cognitive Development 103 (2002) 1–22
OF
22