You are on page 1of 19

This article was downloaded by: [University of Nottingham]

On: 19 March 2015, At: 11:03


Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954
Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,
UK

The Journal of Genetic


Psychology: Research and
Theory on Human Development
Publication details, including instructions for
authors and subscription information:
http://www.tandfonline.com/loi/vgnt20

Communicative Interactions,
Knowledge of a Second
Language, and Theory of Mind
in Young Children
a

George Berguno & Dermot M. Bowler

The American International University Department


of Psychology Richmond London
b

City University Department of Psychology London


Published online: 07 Aug 2010.

To cite this article: George Berguno & Dermot M. Bowler (2004) Communicative
Interactions, Knowledge of a Second Language, and Theory of Mind in Young Children,
The Journal of Genetic Psychology: Research and Theory on Human Development,
165:3, 293-309, DOI: 10.3200/GNTP.165.3.293-309
To link to this article: http://dx.doi.org/10.3200/GNTP.165.3.293-309

PLEASE SCROLL DOWN FOR ARTICLE


Taylor & Francis makes every effort to ensure the accuracy of all the
information (the Content) contained in the publications on our platform.
However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness,
or suitability for any purpose of the Content. Any opinions and views
expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the
Content should not be relied upon and should be independently verified with
primary sources of information. Taylor and Francis shall not be liable for any

losses, actions, claims, proceedings, demands, costs, expenses, damages,


and other liabilities whatsoever or howsoever caused arising directly or
indirectly in connection with, in relation to or arising out of the use of the
Content.

Downloaded by [University of Nottingham] at 11:03 19 March 2015

This article may be used for research, teaching, and private study purposes.
Any substantial or systematic reproduction, redistribution, reselling, loan,
sub-licensing, systematic supply, or distribution in any form to anyone is
expressly forbidden. Terms & Conditions of access and use can be found at
http://www.tandfonline.com/page/terms-and-conditions

The Journal of Genetic Psychology, 2004, 165(3), 293309

Downloaded by [University of Nottingham] at 11:03 19 March 2015

Communicative Interactions, Knowledge


of a Second Language, and Theory
of Mind in Young Children
GEORGE BERGUNO
Department of Psychology
Richmond, The American International University in London
DERMOT M. BOWLER
Department of Psychology
City University, London

ABSTRACT. In this study, the authors explored the effect that particular patterns of communicative interactions may have on young childrens understanding of representations
and the link between knowledge of a 2nd language and theory of mind. The authors tested 140 single language users and 57 dual language users (aged 3-4 years old) on a deceptive task with 3 experimental conditions. In the deceptive-context condition, an experimenter made a reference to the deceptive object. In the older peer condition, an
experimenter made a reference to an older child who was actively participating in the
experimental manipulations. In the deceptive-interaction condition, the experimenter made
reference to his or her own deceptive interventions. The most significant finding was that
knowledge of a 2nd language significantly improved young childrens understanding of
both mental and nonmental representations. Moreover, the results indicated that an experimenters reference to a deceptive interaction greatly facilitated 3-year-old childrens
understanding of false belief.
Key words: second language, deception, theory of mind

RESEARCHERS HAVE SUGGESTED that young childrens understanding of


mind is dependent on a number of factors such as particular patterns of communicative interactions (Surian & Leslie, 1999), the quality of their social interactions (Cutting & Dunn, 1999), and metalinguistic awareness (Doherty & Perner,
1998). Sullivan and Winner (1991) and Chandler and Hala (1994), who reported
that 3-year-olds succeed on false-belief tasks when these tasks are framed in the
Address correspondence to George Berguno, Department of Psychology, Richmond, The
American International University in London, Queens Road, Richmond, Surrey TW10 6JP,
England; bergung@richmond.ac.uk (e-mail).
293

Downloaded by [University of Nottingham] at 11:03 19 March 2015

294

The Journal of Genetic Psychology

context of a deceptive game, have noted that certain patterns of communicative


interactions can facilitate young childrens understanding of mind. It also has
been shown that a small variation in the standard false-belief procedure may lead
to high success rates among 3-year-olds. For example, Lewis (1994) demonstrated that asking children to retell the story before the experimenter asks the test
question facilitates the performance of 3-year-olds on a false-belief task, although
simply repeating the story to the children does not significantly improve the likelihood of correct responses. Similar results were obtained by Mitchell and Lacohee (1991), who were able to improve young childrens performance on a classic deceptive box task by having them select and then post a picture that
represented their own false belief.
In more recent studies, Appleton and Reddy (1996) showed that explaining
particular instances of deception shown on videotaped recordings to young children will aid them on false-belief tasks, whereas Siegal and Peterson (1994)
demonstrated that 3-year-olds can correctly distinguish between lies and mistakes.
A related study of childrens understanding of the appearancereality distinction
by Rice, Koinis, Sullivan, Tager-Flusberg, and Winner (1997) showed that 3-yearolds performed well on the appearancereality test when their aim was to trick
someone or when they did not need to simultaneously hold two conflicting object
identities in mind. Finally, a more controversial example concerns the use of more
explicit forms of test questions. Thus, Siegal and Beattie (1991), followed by Surian and Leslie (1999), found that 3-year-olds performed significantly better on a
false-belief task following the inclusion of the word first in the prediction question, Where will Sally look first for her marble? However, Clements and Perner (1994) failed to replicate those results. Bowler, Briskman, and Grice (1999)
have shown that childrens performance on both a false-drawing task and an unexpected-transfer task improved significantly following an unexpected intervention
by a naughty puppet. Furthermore, Berguno (1999) observed that 3-year-olds performance on a standard false-belief task could be improved if the experimenter
made a reference to a deceptive interaction. However, the introduction of a second child into the experimental area, and to whom the experimenter could make
reference, did not affect childrens scores. It was suggested that the mere presence
of another child was insufficient to motivate them to apply their understanding of
false belief.
Taken as a whole, the studies that we reviewed suggest that experiments carried out using the standard false-belief procedure have seriously underestimated
young childrens understanding of mind. Furthermore, they suggested that young
childrens theory-of-mind abilities could not be studied independently of the
social networks and communicative patterns of interaction in which they manifest. The question is still debated, however, whether young childrens poor falsebelief performance is an artifact of the most common experimental procedures
used in theory-of-mind research or if the developmental changes on these tasks
reflect genuine conceptual change in childrens understanding of mind.

Downloaded by [University of Nottingham] at 11:03 19 March 2015

Berguno & Bowler

295

In an attempt to address the theoretical controversies arising from contradictory findings in theory-of-mind research, Wellman, Cross, and Watson (2001)
carried out a meta-analysis of 178 separate studies of young childrens understanding of false belief. Those researchers found that there was a consistent developmental trend in preschoolers understanding of false belief, such that correct
performance on false-belief tasks significantly increased with age. Moreover,
those results were unaffected by the many task variants or by cultural background.
Thus, it would appear that preschoolers understanding of mind goes through genuine conceptual change. Nevertheless, according to Wellman et al. there was one
factor that was found to interact with agethe temporal marking of test questions enhanced the performance of the older children. Because the temporal clarification of test questions did not significantly improve the younger childrens
performance, Wellman et al. rejected the hypothesis that young childrens less
developed conversational skills did not allow them to reveal their conceptual
understanding of mind (Siegal, 1997; Siegal & Peterson, 1994).
In response to Wellman et al.s (2001) findings, Astington (2001) argued that
more focus needs to be given to the role that language plays in theory-of-mind
development. According to Astington, there are at least two ways in which linguistic ability may be fundamental to the development of childrens understanding of mind. First, language can be used to represent mental states, including false
beliefs. Second, language is also one of the ways by which we become aware of
anothers mental attitude. It therefore becomes important for researchers to conceptualize the development of childrens understanding of mind in a way that
takes into consideration the interaction between childrens internal resources and
their social world.
Evidence for the role of language in childrens understanding of mind has
come from researchers who have examined the association between preschoolers
general language abilities and their performance on theory-of-mind tasks (Astington & Jenkins, 1999; Cutting & Dunn, 1999; Dunn, Brown, Slomkowski,
Tesla, & Youngblade, 1991; Farrar & Maag, 2002; Hughes & Dunn, 1997; Jenkins & Astington, 1996). On the whole, those studies have indicated that childrens
ability to succeed at false-belief tasks is associated with general language abilities. Moreover, childrens understanding of the semantics of mental state terms
develops in parallel with their mastery of the various theory-of-mind tasks
(Moore, Pure, & Furrow, 1990). Yet, although those studies are suggestive of a
causal relationship between language development and the emergence of a
theory of mind, they do not specify the nature of the link.
Harris (1996, 1999) proposed that a key influence on young childrens understanding of false beliefs is their developing ability to sustain a conversation.
Specifically, young childrens emergent understanding of others as epistemic subjects (persons who know and think) is rooted in the context of communication.
Harris (1999) argued that, although young children might at first learn to initiate
communication to mobilize joint goal-directed action, as they grow older they

Downloaded by [University of Nottingham] at 11:03 19 March 2015

296

The Journal of Genetic Psychology

would tend to use conversation as a means of exchanging information and comparing thoughts. According to Harris (1999), there is now good evidence to suggest that this pattern of development in childrens conversational exchanges is
increasingly evident between the ages of 18 and 36 months. For example, both
Snow, Pan, Imbens-Bailey, and Herman (1996) and Morford and GoldinMeadow (1997) demonstrated that young children show an increasing tendency
to refer to objects and events that are not found within their immediate visual
field. Moreover, between the ages of 2 and 3 years, young children also demonstrate an increasing ability to both offer and request clarification of conversational
exchanges. Those findings would suggest that young children come to recognize
that others are not only agents capable of action, but that they are also agents who
form mental representations of their environment and that these representations
may need updating.
A similarly strong argument concerning the relationship between language
and theory of mind was proposed by de Villiers and de Villiers (2000), who argued
that language is a necessary condition for the development of a representational
theory of mind. The crucial concept is that a specific form of linguistic construction, known as sentential complements, makes possible the childs acquisition of a representational theory of mind. Elaborating on the syntax of complementation, de Villiers and de Villiers noted that sentences involving mental states
require embedded propositions (complements), which in turn require the use of
one of three kinds of verbs. Complements may use a verb of desire (such as want),
or a communication verb (such as say), or a mental state verb (such as think or
believe). The crucial characteristic of the syntax of complementation is that,
although the overall sentence may be true, the embedded proposition may be
false. For example, the sentence, John believed that it snowed in Washington on
Monday, may be true, but the embedded complement, It snowed in Washington on Monday, may still be false. Thus, the syntax of complementation allows
one to represent a persons mental state as contradicting reality.
Support for de Villiers and de Villierss (2000) theory of linguistic determinism came from a training study by Lohmann and Tomasello (2003). In that
study, the researchers investigated whether particular linguistic interactions
causally influence the development of false-belief understanding in young children by randomly assigning one hundred thirty eight 3-year-olds to one of four
conditions. In the full-training condition, the deceptive aspects of training objects
were highlighted and the experimenter referred to the objects using sentential
complement syntax. In the discourse-only condition, the deceptive aspects of the
target objects were highlighted without the use of sentential complements or mental state verbs. In the no-language condition, the deceptive aspects of the training
objects were highlighted nonverbally. Finally, in the sentential complement-only
condition, the deceptive aspects of the target objects were not highlighted in any
way, but the experimenter talked about the objects using mental state verbs and
sentential complement syntax. Children in the no-language condition did not

Downloaded by [University of Nottingham] at 11:03 19 March 2015

Berguno & Bowler

297

improve their false-belief understanding. In contrast, children in the discourseonly and sentential complement-only conditions showed improved false-belief
understanding after training. However, the children in the full-training condition
showed the largest training effect.
Hale and Tager-Flusberg (2003) carried out a similar training study investigating the role of language in theory-of-mind development. Sixty 3-year-olds who had
failed false-belief and sentential complement preliminary tests were randomly
assigned to one of three conditions: (a) a false-belief training group, (b) a sentential
complements training group, or (c) a relative clauses training group (which acted as
a control group). To ensure that only one particular linguistic construction was under
investigation (sentential complements), the experimenters restricted themselves to
communication verbs throughout training sessions and made no use of mental state
verbs. Training in false belief improved the childrens theory-of-mind scores; however, the children who were trained on relative clauses were unable to improve their
theory-of-mind understanding. More important, the group trained in sentential complements significantly improved its performance on a variety of theory-of-mind tasks
and acquired new linguistic knowledge.
The two training studies we reviewed provide strong support for the idea that
language plays an active role in influencing childrens understanding of mind.
More specifically, those studies lent support to the linguistic determinism hypothesis proposed by de Villiers and de Villiers (2000), which showed that particular forms of syntactic construction provide the necessary conditions for the development of a representational theory of mind. However, although language may
be a necessary condition for the development of a theory of mind, it may not be
a sufficient condition. Perner, Sprung, Zauner, and Haider (2003) showed that
although the syntax of complementation is obligatory in German for both want
sentences as well as say or think sentences, young German-speaking children
found it easier to remember the complements of desire statements than the complements of communication or mental state sentences. Moreover, Tardiff and
Wellman (2000) found that Mandarin- and Cantonese-speaking Chinese children
used desire terms more often and earlier than they used belief terms, even though
both of these terms use the same relatively simple grammatical constructions.
Thus, the findings of Perner et al. and Tardiff and Wellman suggest that there may
be constraints on the linguistic determinism of theory-of-mind development.
Those previously mentioned ideas motivated the present study, in which we
sought to test a number of new hypotheses. The first and most important question raised in this study is one that is relatively new to theory-of-mind research
does the fluent knowledge of a second language provide a cognitive advantage
on tasks that test a young childs understanding of representations? There have
been few studies of the relationship between theory-of-mind ability and knowledge of a second language, but the rationale for the present study came from a
number of sources. In addition to the training studies that we already mentioned
(Lohmann & Tomasello, 2003; Hale & Tager-Flusberg, 2003), there is evidence

Downloaded by [University of Nottingham] at 11:03 19 March 2015

298

The Journal of Genetic Psychology

to support a link between high levels of language proficiency and an understanding of false belief (Happ, 1993, 1995). Furthermore, a series of four experimental studies by Doherty and Perner (1998) have provided evidence for a
strong link between metalinguistic awareness and false-belief understanding.
They also have suggested that, because their findings confirm that metalinguistic awareness can be demonstrated in 4-year-olds, it may be that there is also a
strong link between metalinguistic awareness and young childrens understanding of representations. On the basis of this argument, one would expect to find
an association between young childrens fluent knowledge of a second language
and their understanding of the appearancereality distinction. Bialystoks (1997)
argument, that children who are bilingual tend to outperform those who are
monolingual on tasks that require high levels of cognitive control, also would
support a strong link between second language knowledge and metarepresentational abilities. Thus, those studies provided the rationale for our present study,
in which we examined the effects of the knowledge and use of a second language
on both young childrens understanding of false belief and their knowledge of
the appearancereality distinction.
Although the primary focus of our study concerned the role of language in
theory-of-mind understanding, there were two additional hypotheses that we
investigated. First, in relationship to the idea that childrens understanding of
mind is related to the quality of their social interaction (Cutting & Dunn, 1999),
we investigated the effect that a second childs active participation in the experimental manipulations has on another childs understanding of a deceptive object.
Second, we also focused on childrens understanding of mind as given within certain communicative contexts (Surian & Leslie, 1999). In particular, we examined
the effect that an experimenters reference to a deceptive object had on a childs
understanding of the appearancereality distinction, as well as their ability to
remember their own previous false belief. We also investigated the possible
effects of an experimenters reference to a deceptive interaction.
Method
Participants
We tested a total of 197 children that we recruited from several preschools
serving the northern half of an inner London borough, of which 140 were single
language users and 57 spoke a second language. We divided the single language
users into two groups: (a) a group of eighty 3-year-olds (M age = 41.8 months,
SD = 2.8) and (b) a group of sixty 4-year-olds (M age = 53 months, SD = 3). We
similarly divided the children who spoke a second language into two groups: (a)
a group of thirty-one 3-year-olds (M age = 43.2 months, SD = 1.9) and (b) a group
of twenty-six 4-year-olds (M age = 52.5 months, SD = 3). There were approximately an equal number of girls and boys across both age groups and conditions,

Berguno & Bowler

299

Downloaded by [University of Nottingham] at 11:03 19 March 2015

TABLE 1. Distribution of Gender Across Conditions

Condition

3-year-olds (n = 80)
Boys
Girls

4-year-olds (n = 60)
Boys
Girls

Standard
Context
Older peer
Interaction

13 (3)
16 (5)
14 (4)
14 (4)

13 (4)
10 (3)
11 (3)
13 (4)

15 (5)
12 (3)
14 (4)
13 (3)

12 (4)
10 (4)
9 (2)
8 (2)

Note. Numbers in parentheses refer to proportion of children with knowledge of a second


language.

as illustrated in Table 1. We tested all of the children individually in a separate,


quiet area of their school. Parental consent was obtained in writing.
Children were categorized as dual language users if the school staff members could confirm that the child used English in the school setting and if parents
also could confirm that the child preferred to use a language other than English
at home. Single language users were children who, according to parental reports,
had little or no knowledge of any language other than English.
Materials
In the testing area at the preschool, we placed a small transparent, plastic
container filled with water to serve as an aquarium. We placed seashells, pebbles,
and seaweed inside the container. We used a bright pen in the shape of a fish, with
round eyes, shiny red and yellow scales, and a removable tail as the deceptive
object.
Procedure
Each child was randomly assigned within age groups to one of four conditions:
the deceptive-context condition (n = 48), the older-peer condition (n = 48), the
deceptive-interaction condition (n = 48), or the standard-task condition (n = 53).
In the standard-task condition, we took each child to a separate room and
showed them the aquarium filled with water and the target objecta pen that looked
like a bright red fish. The experimenter pointed to the tricky object and said: Look
at this. What does this look like to you? If the child replied by saying it looked
like a fish, the experimenter responded, Yes, it looks like a fish. If the child did
not know what it looked like or did not respond, the question was repeated: Look
again. Look very closely. What does this look like to you? The child was required
to say what the target object looked like before the experiment could continue.

Downloaded by [University of Nottingham] at 11:03 19 March 2015

300

The Journal of Genetic Psychology

After the child had given the correct answer, the experimenter took the fish-like
object out of the water and pulled its tail off to reveal a pen. The experimenter then
proceeded to write a word or two on a pad of paper located on a nearby table. Next,
the experimenter offered the pen to the child and said, Did you see what I did? Here,
you try. Try and write something with this. After the child had written something,
the experimenter asked, So, what is it really and truly? The child now had to correctly identify the deceptive object before the experiment could continue. If the child
did not respond, the question was repeated, but an incorrect answer meant that the
testing was discontinued and the child was not included in the study.
Next, the experimenter replaced the tail onto the pen so that it would look
like a fish. The experimenter returned the fish-like object to the aquarium and
asked the child the three test questions.
The Questions
Reality question. Now, for real, what is this really and truly?
Appearance question. Now, when you look at this, what does it look like?
False-belief-for-self question. When I first showed this to you, what did you
think it was?
To pass the appearancereality task, we required the child to give correct
answers to both the appearance and reality questions. Questions were counterbalanced with six different orders of presentation, but children were not presented the
questions in forced choice format, because the order of response choices also
would have had to be counterbalanced across questions.
We designed the deceptive-context condition exactly like the appearance
reality task, with the exception that the experimenter preceded the test questions by
making a reference to the trick object and the situation as a whole, This is a tricky
situation, and this is a tricky object because it looks like one thing but it really is
another.
We designed the older-peer condition task exactly as we did the standard-task
condition, with the exception that a 5-year-old child accomplished all the manipulations, whereas the experimenter asked the questions. This set-up allowed the experimenter the opportunity to ask the false-belief question in reference to the childs
older peer: When Danielle first showed this to you, what did you think it was?
In the deceptive-interaction condition, we asked, You see, Im a tricky person, because when you came in I played a trick on you, showing you this toy that
looks like one thing but it really is another.
Results
We conducted two main analyses. First, we ran a hierarchical 4 2 2 2
(Condition Age False Belief for Self Language) loglinear analysis, which

Downloaded by [University of Nottingham] at 11:03 19 March 2015

Berguno & Bowler

301

focused on the childrens performance on the false-belief task. Second, we ran a


separate hierarchical 4 2 2 2 (Condition Age Appearance Reality Language) loglinear analysis, which examined childrens understanding of the
appearancereality distinction.
The loglinear analyses revealed an interaction between language and performance on the false-belief-for-self task, likelihood ratio (LR): 2(1, N = 197) = 6.57,
p = .01, with the dual language users outperforming the single language users. This
result is shown in Table 2, which shows that of 140 single language users, 76 (54%)
passed the false-belief-for-self question. The 57 children who had knowledge of a
second language, in contrast, showed a better pattern of responding, with 42 (74%)
children giving correct responses to the false-belief-for-self task.
The loglinear analysis also revealed an interaction between language and childrens scores on the appearancereality task, LR: 2(1, N = 197) = 6.715, p < .01,
with the dual language users performing better than the single language users. Table
3 shows that of 140 single language users, 73 passed the appearancereality task.
However, 41 of the 57 children with knowledge of a second language passed.
Expressed as percentages, only 52% of the single language users responded correctly to these test questions, compared with 72% correct responses for the dual
language users.
Thus, our results showed that, across both tasks, the dual language users
showed better rates of correctly responding than did the single language users.
However, to test for the possibility that the effect for language on both tasks might

TABLE 2. False Belief for Self Language Crosstabulation


Language
False belief
Fail
Pass

Single language users

Dual language users

64
76

15
42

TABLE 3. Appearance-Reality Task Language Crosstabulation

Appearance-reality task
Fail
Pass

Language
Single language users
Dual language users
67
73

16
41

Downloaded by [University of Nottingham] at 11:03 19 March 2015

302

The Journal of Genetic Psychology

have been influenced by age, we carried out separate chi-square analyses for the
single and dual language users.
With regard to the performance of the single language users, separate chi-square
analyses revealed a significant association between age and performance on both the
false-belief-for-self task, 2(1, N = 140) = 10.45, p > .002, and the appearance
reality task, 2(1, N = 140) = 6.96, p > .009, with the older children showing the better rates of correct responding. Thus, of 80 single language 3-year-olds, only 34
(43%) gave correct responses to the appearancereality question, whereas from the
total of 60 single language 4-year-olds, 39 (65%) passed this task. Of the 80 single
language 3-year-olds, 34 (43%) passed the false-belief-for-self task, whereas 42
(70%) of the 60 single language 4-year-olds succeeded at this task.
With regard to the performance of the dual language users, separate chi-square
analyses indicated that there was no significant association between age and performance across both tasks, 2(1, N = 57) = 0.01, p > .93, for the false-belief-forself task, and 2(1, N = 57) = 0.596, p > .45, for the appearancereality task. Thus,
of 31 dual language 3-year-olds, 21 (68%) passed the appearancereality task. This
result compared favorably with the performance of the 26 dual language 4-yearolds, of whom 20 (77%) gave correct responses to this question. Of the 31 dual language 3-year-olds, 23 (74%) gave correct responses to the false-belief-for-self question, a result that compares well with the performance of the 26 dual language
4-year-olds, of whom 19 (73%) passed this task.
Thus, the single language users exhibited the usual age-related changes in performance on both tasks, with the older children producing the better scores. In contrast, the younger dual language users showed a pattern of correctly responding
across both tasks, which was comparable to their older counterparts. These findings support the view that it was the younger childrens knowledge of a second
language that facilitated their performance on both tasks. Moreover, it would
appear that the effect of language was greater for the false-belief-for-self task.
For the effects of conditions on the two tasks, the loglinear analyses showed
that there were no significant differences across conditions for all children on the
false-belief-for-self task, LR: 2(3, N = 197) = 4.28, p < .24. However, a close
inspection of the data showed a strong ceiling effect for the 4-year-olds, which
prompted us to run a separate 2 4 (Condition False Belief for Self) chi-square
analysis for the 3-year-olds. This analysis revealed a significant interaction
between condition and performance on the false-belief-for-self question, 2(3, N
= 111) = 9.83616, p < .03, with children in the deceptive interaction condition
showing the best scores. Further examination of the data showed that, although
children in all three experimental conditions showed improved scores relative to
the standard task condition, childrens understanding of false belief for self was
best facilitated by an experimenters reference to a deceptive interaction. Table 4
shows these results, and it can be observed that of 28 children in the control group,
only 8 (29%) gave correct responses to this test question. In contrast, of 27 children in the deceptive-interaction condition, 19 (70%) responded correctly.

Berguno & Bowler

303

TABLE 4. Condition False-Belief-for-Self Crosstabulation


Condition
False belief

Downloaded by [University of Nottingham] at 11:03 19 March 2015

Fail
Pass

Standard

Deceptive
context

Older peer

Deceptive
interaction

20
8

13
15

13
15

8
19

The analyses also showed that there were no significant differences across
conditions for all children on the appearancereality task, LR: 2(3, N = 197) =
2.35, p < .51. Because the data indicated that the 4-year-olds were performing at
ceiling, we performed a separate 2 4 (Condition False Belief for Self) chisquare analysis for the 3-year-olds, but this analysis did not reveal an interaction
between condition and performance on the appearancereality task, 2(3, N =
111) = 3.35, p < .35.
For the effects of age on performance across the two tasks, the analyses
revealed an interaction between age and performance on the false-belief-for-self
question, LR: 2(1, N = 197) = 7.85, p < .006, with the older group performing
better than the younger one. Thus, reviewing 111 of the 3-year-olds, only 57 (51%)
passed the false-belief-for-self task. In contrast, reviewing 86 of the 4-year-olds,
61 (71%) responded correctly to this test question. Our analyses also revealed an
interaction between age and performance on the appearancereality task, LR: 2(1,
N = 197) = 7.31, p < .007, with the older group performing better than the younger
one. Of one hundred eleven 3-year-olds, only 55 (50%) passed the
appearancereality task. This result contrasts with the performance of the eightysix 4-year-olds, of whom 59 (69%) responded correctly to this task.
Finally, a separate hierarchical 4 2 2 6 (Condition False Belief for Self
Appearance Reality Order) loglinear analysis indicated that there were no order
effects on either the false-belief-for-self task, LR: 2(5, N = 197) = 2.54, p < .77, or the
appearancereality task, LR: 2(5, N = 197) = 1.35, p < .93. This analysis also showed
that there was a significant association between childrens performance on the
false-belief-for-self task and on the appearancereality task, LR: 2(1, N = 197) =
49.312, p < .0001. A close inspection of the data revealed that of 114 children who
passed the appearancereality task, 92 (81%) also responded correctly to the falsebelief-for-self task. Moreover, of 83 children who failed the appearancereality
task, 57 (69%) also failed the false-belief-for-self task.
Discussion
The most important finding of our study was that the knowledge of a second
language made a significant difference to childrens performances on both the

Downloaded by [University of Nottingham] at 11:03 19 March 2015

304

The Journal of Genetic Psychology

appearancereality and the false-belief-for-self tasks. Although in this study we


did not claim that the dual language users were bilingual, this finding is consistent with the research results documented in the developmental literature on bilingualism, which have suggested that there are definite cognitive advantages to
learning two languages at an early age. In the present study, we were not able to
test the interaction between the effects of a reference to a deceptive interaction
and knowledge of a second language because of the requirement of a minimum
number of expected values in loglinear analyses. We have planned future studies
to address this question using larger samples.
Researchers have reported that bilinguals score higher than do monolinguals
on tests of language proficiency, concept formation, and nonverbal intelligence
(Bialystok, 1986, 1988; Diaz, 1985). Recent findings show that bilingual speakers outperform monolingual speakers on tests of metalinguistic abilities (Bialystok, 1997; Galambos & Goldin-Meadow, 1990). Bialystok (1988) has pointed
out that a consistent finding throughout research on bilingualism is that bilingual
speakers perform better than do monolingual speakers of the same age on tasks
that demand high levels of control (executive function), but that there is no advantage to being bilingual on tasks that simply require an analysis of representational
structures. At first sight then, it would appear that there is no reason why the dual
language users in our study should have performed better than did the single language users, until we remembered that both the appearancereality and the falsebelief-for-self task required metarepresentational abilities. According to our findings, it would appear that the metacognitive abilities of dual language users would
extend beyond the domain of linguistic processing. Precisely why bilinguals are
better able to resist distraction has yet to be determined (Bialystok, 1999), and
exactly why bilinguals or dual language users should be better at theory-of-mind
tasks is a question to be resolved by future research. It could be argued that knowledge of a second language confers the advantage of having at least two linguistic representations to every referent, and that this metalinguistic ability affects
theory-of-mind abilities (Charman & Schmueli-Goetz, 1998; Doherty & Perner,
1998; Cutting & Dunn, 1999). One could argue that being a fluent speaker of two
languages allows one to participate in two social worlds. The dual language child
could then be expected to be more attuned and sensitive to subtleties of communicative interactions. These conclusions, on the basis of our observations, serve
to constrain linguistic determinist accounts such as those of de Villiers and de Villiers (2000) and support studies by Tardiff and Wellman (2000) and Perner et al.
(2003), who have concluded that the structurallinguistic aspects of any single
language do not adequately specify theory-of-mind competence.
Because there is now evidence to suggest that there is a link between linguistic proficiency and an understanding of false belief (Happ, 1993, 1995), the
high rates of success among the dual language users on the false-belief-for-self
task might have been because of high levels of language proficiency rather than
knowledge of a second language per se. However, an alternative explanation has

Downloaded by [University of Nottingham] at 11:03 19 March 2015

Berguno & Bowler

305

been presented by Doherty and Perner (1998), who were able to provide evidence
across four studies that it is metalinguistic awareness, rather than general language proficiency, that is related to the ability to understand the representational
nature of mind. In their first experiment, Doherty and Perner tested 3- and 4-yearolds on a standard version of the unexpected transfer task, as well as a synonymjudgment task. The results indicated that childrens performance on the two tasks
was highly correlated, with the understanding of false-belief and metalinguistic
abilities emerging at about age 4. However, because it was clearly possible that
the two tasks were strongly associated because of an underlying dependency on
general verbal intelligence, the researchers proceeded to run a second experiment
in which children were given the British Picture Vocabulary Test along with the
standard false-belief task and the synonym judgment task. That second experiment replicated the essential findings of Experiment 1, and demonstrated that
there was a strong association between success on the false-belief and synonym
tasks, even after performance on the British Picture Vocabulary Test had been partialled out. Thus, the association between the two tasks could not be explained by
general verbal intelligence.
In their third experiment, Doherty and Perner (1998) explored the possibility
that an early metalinguistic understanding might have been obscured by the
demands of the synonym task. The researchers therefore devised a simplified version of this task. Even so, childrens abilities to produce synonyms and to understand false belief were strongly correlated. In Experiment 4, Doherty and Perner
examined the possibility that young childrens poor performances on the synonym
task were simply because of an unwillingness on the part of the children to think
of a different answer to the target word. Those researchers tested that idea by introducing a control task in which children were required to misname a target object
by pretending it was something else. However, childrens performances on this
pretend control task were significantly better than were their performances on the
synonym-judgment task, thus demonstrating that childrens difficulties on the latter task were not simply because of a reluctance to provide an alternative description of an object. Furthermore, childrens rates of success on the synonym and
false-belief tasks were strongly associated. Thus, Doherty and Perner were able to
show across four experiments that there is a close relationship between metalinguistic awareness and false-belief understanding, an association that cannot be
explained by an appeal to general verbal intelligence. The findings reported in the
present article are consistent with Doherty and Perners idea that there is an underlying conceptual basis to metalinguistic awareness and false-belief understanding,
given that the improved performance of the dual language children extended to the
appearancereality task and that there was a strong association between childrens
performance on the two tasks, as was discussed previously.
Consistent with previous research, in our present study, we found significant differences in performance between age groups on both the
appearancereality and the false-belief-for-self tasks, with the older children

Downloaded by [University of Nottingham] at 11:03 19 March 2015

306

The Journal of Genetic Psychology

obtaining better scores than the younger ones. However, the introduction of a
second child, who actively participated in the experimental manipulations, did
not improve the scores of the children tested. The 5-year-old child who was
used in this condition had been trained by the researcher to carry out the expertask condition. Although we hypothesized that the presence of a second child
would improve task performance, in retrospect, the lack of effect found here is
perhaps not surprising. This result occurred because there turned out to be few
spontaneous interactions between children in the older peer condition. Furthermore, a setting where one child pretends to be an experimenter is an unusual one, which does not in any way resemble the interactive patterns to be found
in childrens natural play environments.
Moreover, in our study, we found that the experimenters reference to a
deceptive object did not significantly improve childrens scores on either the
appearancereality or the false-belief-for-self questions. The experimenters reference to a deceptive interaction also did not facilitate childrens understanding
of the appearancereality distinction. Nevertheless, the experimenters reference
to a deceptive interaction significantly improved the performance of the 3-yearolds on the false-belief-for-self question. On the basis of these findings, we conclude and propose an explanation. We conclude that 3-year-olds have a developing understanding of false belief, but one which shows itself in certain contexts
only. Furthermore, we proposed that if early theory-of-mind abilities are embedded in particular patterns of communicative interactions, then a communicative
exchange between experimenter and child, which highlights these interactions in
the form of a metacommunication, may well enhance childrens performance on
a theory-of-mind task. This line of reasoning is supported by a number of recent
studies (Chandler, Fritz, & Hala, 1989; Dalke, 1995; Lewis, 1994; Siegal & Peterson, 1994).
Our finding that neither a reference to a deceptive object nor a reference to
deceptive interaction improved the 3-year-olds understanding of the appearance
reality distinction would appear to be in conflict with the findings reported by
Rice et al. (1997). However, the deceptive context that Rice et al. introduced into
their experiment was a complex interaction of a reference to a deceptive interaction, a reference to a deceptive object, and an encouragement to the test child to
deceive another child. It is therefore not at all clear which aspect or combination
of these aspects was responsible for the improved scores of the 3-year-olds on
their appearancereality task. Our results suggest that a reference to a deceptive
interaction is insufficient on its own to improve young childrens performances
on this task.
A number of questions motivated the present study. First, and most important,
would knowledge of a second language provide a cognitive advantage on a test of
false belief for self and the appearancereality distinction? Second, would the
active participation of another child facilitate childrens understanding of false
belief or their understanding of the appearancereality distinction? Third, what

Berguno & Bowler

307

Downloaded by [University of Nottingham] at 11:03 19 March 2015

kinds of referencing might improve childrens performances on the false-belieffor-self and appearancereality tasks? In this study, we examined two kinds of referencing: (a) an experimenter reference to a deceptive object and (b) an experimenter reference to a deceptive interaction. Our findings indicate that the fluent
knowledge of a second language significantly improves childrens understanding
of both mental and nonmental representations. In contrast, the active participation
of a second child does not affect childrens performances on the two set tasks. With
respect to the different kinds of referencing, only an experimenter reference to a
deceptive interaction facilitates 3-year-olds understanding of false belief for self.
REFERENCES
Appleton, M., & Reddy, V. (1996). Teaching 3-year-olds to pass false belief tests: A conversational approach. Social Development, 5, 275291.
Astington, J. W. (2001). The future of theory-of-mind research: Understanding motivational states, the role of language, and real-world consequences. Child Development, 72,
685687.
Astington, J. W., & Jenkins, J. (1999). A longitudinal study of the relation between language and theory of mind development. Developmental Psychology, 35, 13111320.
Berguno, G. (1999, April). A reference to a deceptive interaction facilitates young childrens understanding of false belief. Poster presented at the Society for Research in
Child Development Biennial Conference, Albuquerque, NM.
Bialystok, E. (1986). Factors in the growth of linguistic awareness. Child Development,
57, 498510.
Bialystok, E. (1988). Levels of bilingualism and levels of metalinguistic awareness. Developmental Psychology, 24, 560567.
Bialystok, E. (1997). Effects of bilingualism and biliteracy on childrens emerging concepts of print. Developmental Psychology, 33, 429440.
Bialystok, E. (1999). Cognitive complexity and attentional control in the bilingual mind.
Child Development, 70, 636644.
Bowler, D. M., Briskman, J. A., & Grice, S. (1999). Experimenter effects on childrens
understanding of false drawings and false beliefs. The Journal of Genetic Psychology,
160, 443460.
Chandler, M., & Hala, S. (1994). The role of personal involvement in the assessment of
early false-belief skills. In C. Lewis and P. Mitchell (Eds.), Childrens early understanding of mind: Origins and development (pp. 403425). Hove, England: Erlbaum.
Chandler, M. J., Fritz, A. S., & Hala, S. M. (1989). Small scale deceit: Deception as a
marker of 2-, 3- and 4-year-olds early theories of mind. Child Development, 60,
12631277.
Charman, T., & Schmueli-Goetz, Y. (1998). The relationship between theory of mind, language and narrative discourse: An experimental study. Current Psychology of Cognition, 17, 245271.
Clements, W. A., & Perner, J. (1994). Implicit understanding of belief. Cognitive Development, 9, 377395.
Cutting, A. L., & Dunn, J. (1999). Theory of mind, emotion understanding, language, and
family background: Individual differences and interrelations. Child Development, 70,
853865.
Dalke, D. E. (1995). Explaining young childrens difficulty on the false-belief task: Representational deficits or context-sensitive knowledge? British Journal of Developmental

Downloaded by [University of Nottingham] at 11:03 19 March 2015

308

The Journal of Genetic Psychology

Psychology, 13, 209222.


de Villiers, J., & de Villiers, P. (2000). Linguistic determinism and the understanding of
false beliefs. In P. Mitchell & K. J. Riggs (Eds.), Childrens reasoning and the mind (pp.
191228). Hove, England: Psychology Press.
Diaz, R. M. (1985). Bilingual cognitive development: Addressing three gaps in recent
research. Child Development, 56, 13761388.
Doherty, M., & Perner, J. (1998). Metalinguistic awareness and theory of mind: Just two
words for the same thing? Cognitive Development, 13, 279305.
Dunn, J., Brown, J., Slomkowski, C., Tesla, C., & Youngblade, L. (1991). Young childrens
understanding of other peoples feelings and beliefs: Individual differences and their
antecedents. Child Development, 62, 13521366.
Farrar, M. J., & Maag, L. (2002). Early language development and the emergence of a theory of mind. First Language, 22, 197213.
Galambos, S. J., & Goldin-Meadow, S. (1990). The effects of learning two languages on
levels of metalinguistic awareness. Cognition, 34, 156.
Hale, C. M., & Tager-Flusberg, H. (2003). The influence of language on theory of mind:
A training study. Developmental Science, 6, 346359.
Happ, F. G. E. (1993). Communicative competence and theory of mind in autism: A test
of relevance theory. Cognition, 48, 101119.
Happ, F. G. E. (1995). The role of age and verbal ability in the theory-of-mind task performance of subjects with autism. Child Development, 66, 843855.
Harris, P. (1996). Desires, beliefs, and language. In P. Carruthers & P. K. Smith (Eds.),
Theories of theories of mind (pp. 200220). Cambridge, MA: Cambridge University
Press.
Harris, P. (1999). Acquiring the art of conversation: Childrens developing conception of
their conversational partner. In M. Bennett (Ed.), Developmental psychology: Achievements and prospects. Philadelphia: Psychology Press.
Hughes, C., & Dunn, J. (1997). Pretend you didnt know: Young childrens talk about
mental states in pretend play. Cognitive Development, 12, 477499.
Jenkins, J., & Astington, J. W. (1996). Cognitive factors and family structure associated
with theory of mind development in young children. Developmental Psychology, 32,
7078.
Lewis, C. (1994). Episodes, events and narratives in the childs understanding of mind. In
C. Lewis & P. Mitchell (Eds.), Childrens early understanding of mind: Origins and
development (pp. 457480). Hove, England: Erlbaum.
Lohmann, H., & Tomasello, M. (2003). The role of language in the development of falsebelief understanding: A training study. Child Development, 74, 11301144.
Mitchell, P., & Lacohee, H. (1991). Childrens early understanding of false belief. Cognition, 39, 107127.
Moore, C., Pure, K., & Furrow, P. (1990). Childrens understanding of the modal expression of certainty and uncertainty and its relation to the development of a representational
theory of mind. Child Development, 61, 722730.
Morford, J. P., & Goldin-Meadow, S. (1997). From here and now to there and then: The
development of displaced reference in Homesign and English. Child Development, 68,
420435.
Perner, J., Sprung, M., Zauner, P., & Haider, H. (2003). Want that is understood well before
say that, think that, and false belief: A test of de Villierss linguistic determinism on
German-speaking children. Child Development, 74, 179188.
Rice, C., Koinis, D., Sullivan, K., Tager-Flusberg, H., & Winner, E. (1997). When 3-yearolds pass the appearancereality test. Developmental Psychology, 33, 5461.
Siegal, M. (1997). Knowing children: Experiments in conversation and cognition (2nd

Downloaded by [University of Nottingham] at 11:03 19 March 2015

Berguno & Bowler

309

ed.). Hove, England: Erlbaum.


Siegal, M., & Beattie, K. (1991). Where to look first for childrens knowledge of false
beliefs. Cognition, 38, 112.
Siegal, M., & Peterson, C. C. (1994). Childrens theory of mind and the conversational territory of cognitive development. In C. Lewis & P. Mitchell (Eds.), Childrens early understanding of mind: Origins and development (pp. 427455). Hove, England: Erlbaum.
Snow, C. E., Pan, B. A., Imbens-Bailey, A., & Herman, J. (1996). Learning how to say
what one means: A longitudinal study of childrens speech act use. Social Development,
5, 5684.
Sullivan, K., & Winner, E. (1991). When children understand ignorance, false belief and
representational change. British Journal of Developmental Psychology, 9, 159171.
Surian, L., & Leslie, A. (1999). Competence and performance in false-belief understanding: A comparison of autistic and normal 3-year-old children. British Journal of Developmental Psychology, 17, 141155.
Tardiff, T., & Wellman, H. M. (2000). Acquisition of mental state language in Mandarinand Cantonese-speaking children. Developmental Psychology, 36, 2543.
Wellman, H. M., Cross, D., & Watson, J. (2001). Meta-analysis of theory-of-mind development: The truth about false belief. Child Development, 72, 655684.

Received December 1, 2003