Downloaded from http://rstb.royalsocietypublishing.org/ on July 28, 2017
rstb.royalsocietypublishing.org
Planning my actions to accommodate
yours: joint action development during
early childhood
Marlene Meyer1, Robrecht P. R. D. van der Wel2 and Sabine Hunnius1
Research
Cite this article: Meyer M, van der Wel RPRD,
Hunnius S. 2016 Planning my actions to
accommodate yours: joint action development
during early childhood. Phil. Trans. R. Soc. B
371: 20150371.
http://dx.doi.org/10.1098/rstb.2015.0371
Accepted: 8 February 2016
One contribution of 15 to a theme issue
‘Attending to and neglecting people’.
Subject Areas:
cognition
Keywords:
joint action, development, early childhood,
action planning
Author for correspondence:
Marlene Meyer
e-mail: [email protected]
1
2
Donders Institute for Brain, Cognition and Behaviour, Radboud University, Nijmegen, The Netherlands
Department of Psychology, Rutgers University, Camden, NJ, USA
SH, 0000-0001-5510-4032
The planning and adjusting of one’s actions in relation to an action partner is
fundamental to smooth joint action. During their first years of life, children
gradually become more engaged in joint actions. Here, we investigated
whether and at what age children take their partner into account in their
action plans to accommodate the other’s actions. We focused on children’s
proactive planning (without prior experience) and flexible adjustment of
action plans over time. In a behavioural study, we tested 96 children from
four age groups (212, 3, 312 and 5 years) in a joint cup-stacking task. Children
passed cups to their partner who had only one hand available (alternating
over time) to build a tower. Children’s response choices were assessed (i.e. passing the cup on the free or occupied side to their partner). The study yielded
two major findings. At all ages, children proactively planned their actions in
a way that accommodated their partner’s actions. However, only by 312 years
did children start to flexibly integrate their partner into their action plans.
Even at age 5, children only showed minimal adjustments to their action partner. Candidate processes underlying these developmental changes (e.g.
inhibitory control, cognitive flexibility, perspective taking) are discussed.
1. Introduction
A fundamental aspect of social interaction is the ability to coordinate our actions
with those of others. In such joint actions, adults plan and adjust their own actions
to accommodate their action partner, and thereby reach a smooth coordination
with the other person. During the first years of life, children gradually become
more engaged in joint actions. Yet initially, their eager contribution displays
little signs of planning ahead for and adjusting to their action partner. When children’s active engagement turns into a smooth coordination in which they plan
ahead to accommodate the other person remains an open question.
(a) Accommodating the other’s actions: evidence from adult joint
actions
Electronic supplementary material is available
at http://dx.doi.org/10.1098/rstb.2015.0371 or
via http://rspb.royalsocietypublishing.org.
In everyday life, adults usually plan and adjust their own actions to accommodate
the actions of others. For instance, when paying for our groceries at the market the
seller might be holding a bag of fruit for us in his left hand. To pay him, we readily
pass the money to his free right hand (instead of his occupied left hand). Besides
such anecdotal evidence, systematic investigations provide empirical evidence
that adults adjust their actions to benefit an interaction partner, and thereby
make the joint action progress more smoothly (e.g. [1–3]). Ray & Welsh [1], for
example, instructed participants to pass a jug of water to another person who
would subsequently use it (e.g. to pour water into a glass). They found that
about 90% of the participants passed the jug by holding the jug’s body such
that the handle was available for the other person. Instead, when participants
were acting alone, they consistently grasped the jug by the handle rather than
& 2016 The Author(s) Published by the Royal Society. All rights reserved.
Downloaded from http://rstb.royalsocietypublishing.org/ on July 28, 2017
the body to perform the action. Adult participants thus
adjusted their actions to accommodate their action partner.
(c) Accommodating the other’s actions: asymmetry in
early childhood
Still, smooth coordination seems to be a challenge in early
childhood. Children have major difficulties in considering the
influence their own actions have on others and in acting accordingly, while being well aware of how others’ actions affect
themselves and the interaction (see [9] for a review). For
instance, if action partners fail to fulfil their part of an interaction, infants and toddlers overtly protest or try to reengage
their partner through social bidding [10,11]. This can already
be observed in early face-to-face interactions. On the other
hand, even beyond infancy, children readily leave their action
partner for other activities [12,13]. This suggests an asymmetry
in development: whereas young children may take into
account how the others’ actions relate to their own actions
and goals, the extent to which they plan and adjust their actions
with respect to how they affect others’ actions is still limited.
Similar asymmetries can be observed in various domains of
social development. For instance, Aureli & Presaghi [14] found
an asymmetric pattern in their longitudinal study on social
engagement in joint play. Observations of mother–child interactions showed that joint play gradually developed from
unilateral engagement by the mother to a more symmetric
and reciprocal engagement of mother and child between
infants’ first and second year of life [14]. Moreover, it has
been suggested that in early joint actions, infants might
represent mainly their own task and goal while only later
representing also their action partner’s task [15–17]. As
alluded to before, by 18 months of age toddlers protest when
their play partner stops his or her contribution to a joint
game (e.g. [11]), whereas it is up to and beyond their third
year of life that children readily leave an ongoing joint action
for a more attractive game [12] or a reward [13]. With respect
to such commitments to a joint action, Hamann and colleagues
suggest that a ‘you-to-me’ commitment develops before a reciprocal commitment (i.e. you-to-me and me-to-you) in the third
year of life [13].
Together, previous developmental research indicates a
major transition in early childhood towards a more reciprocal
engagement in joint actions. Apart from this evidence on
getting and staying engaged in joint actions, there is little
(d) The current study: hypotheses and design
In a behavioural experiment, we investigated whether and at
what age children accommodate an action partner’s actions in
their own action planning. More specifically, we focused on
two aspects: (i) children’s proactive planning, in the absence
of prior experience with how their actions affect those of their
partner; and (ii) children’s flexible adjustment of their action
plans throughout the course of a joint action integrating
changing constraints of their partner’s actions.
The distinction between proactive planning and flexible
adjustment of action plans is important, because they reflect
different psychological and brain processes. The ability to
proactively plan without having prior experience with a particular situation requires taking the action partner’s task and
perspective into account, and to make action predictions for
the action partner based on that information. In terms of
neural activity, this probably involves ( pre)motor cortices
associated with action prediction (e.g. [19,20]) in combination
with the temporo-parietal junction associated with perspective
taking [21]. Proactive planning does not require the ability to
task switch, as there is no preceding task to switch from yet.
Flexible adjustments over the course of trials instead link to a
sensitivity to observing another actor’s discomfort and the
ability to modify one’s behaviour because of it. Thus, flexibly
adjusting action plans may rely on a combination of action
perception (associated with the action observation network including the premotor and supplementary motor
cortex, inferior parietal lobule and superior temporal sulcus,
e.g. [22]) and task-switching mechanisms (associated with
the superior prefrontal and posterior parietal cortex, [23]).
Based on previous findings, we expected that both proactive planning and the flexible adjustment to accommodate a
joint-action partner would improve between the ages of 212
and 5 years. Developmental research suggests that compared
with 212-year-old children, 3-year-olds are significantly better
at coordinating their actions with a joint-action partner [24],
and demonstrate first signs of acknowledging commitment
to a joint action [12]. Moreover, young children’s individual
planning skills improve significantly between the ages of 3
and 5 years (see [25] for a review). It is also around the
same age that children show increasingly more cognitive
flexibility and inhibitory control [26,27], both prerequisites
for inhibiting prepotent responses and adjusting their own
plans to another person. In accordance with this, it was
suggested that children’s planning skills and their inhibitory
control are closely linked [25]. Although representing distinct
aspects of planning, proactive planning and the flexible
adjustment of plans to accommodate a joint-action partner
might thus show a close interaction.
However, we also expected differences between the two
measures. Previous findings suggest that anticipating the
effect one’s own actions have for another person and proactively planning one’s own actions accordingly might be
Phil. Trans. R. Soc. B 371: 20150371
Apart from accomplishing observable joint action goals
through smooth coordination like passing a jug to serve
water, coordinating actions with others also has the potential
to establish social bonds. A growing body of evidence shows
links between fluent interpersonal coordination and affiliation,
both in children [4,5] and adults (see [6] for a review). More
specifically, smooth action coordination between people
makes them feel more connected to each other [7,8]. Establishing affiliation thus is an integral part of social interaction in
addition to achieving observable goals together. During early
interactions with children, as in proto-conversation and peekaboo games, observable goals are often less prominent [9].
This feature of early social interactions may make the role of
affiliation even more pronounced [9].
2
rstb.royalsocietypublishing.org
(b) Accommodating the other’s actions: effects on
affiliation
research on the developmental trajectory of when children
plan and adjust their actions to accommodate their action
partner during a joint action itself. Planning and adjusting
one’s action plan to accommodate another person would
require young children to take into account how their actions
affect their joint-action partner. This goes beyond using their
partner as means to achieve their own goal and thus beyond
what is called using the other as ‘social tool’ (e.g. [18]).
Downloaded from http://rstb.royalsocietypublishing.org/ on July 28, 2017
(a) Participants
In this study, we tested 96 children from four age groups (212, 3, 312
and 5 years). The final dataset consisted of 89 children: 212-year-olds
22 (mean age: 31.5 months, ranging from 30 to 32 months; 11 girls),
3-year-olds 23 (mean age: 36.9 months, ranging from 36 to 38
months; 16 girls), 312-year-olds 22 (mean age: 42.5 months, ranging
from 41 to 43 months; 13 girls), and 5-year-olds 22 (mean age: 60.7
months, ranging from 60 to 61 months; 10 girls). Seven children
were excluded because of parental interference (N ¼ 2), technical
failure (N ¼ 1), experimenter error (N ¼ 1), because of invalid
data (i.e. walking around the table, see Data processing section;
N ¼ 1), or unwillingness to participate (N ¼ 2). Children were
recruited through a database of the Baby Research Center
Nijmegen. This database contains families from Nijmegen
(a medium-sized city in the Netherlands) and the surrounding
areas who volunteer for participation in developmental research.
Parents gave written consent for their child’s enrolment in the
study and stayed with their child throughout the laboratory visit.
The laboratory visit (including debriefing) took about 30 min
and was compensated with either 10 EUR or a children’s book.
(b) Experimental design and materials
A simple joint action task, suitable for all tested age groups, was
developed. The task involved two people jointly stacking cups:
one person passed cups to a second person, who subsequently
stacked the cups to a tower. By means of a custom-made
wooden apparatus (height 15 cm, width 74 cm, depth 60 cm)
with a hidden turning disc (i.e. a manual carousel-type machine),
cups could be presented to the first person one by one in a centred
fashion (figure 1). The apparatus was placed on a table and the
(c) Procedure
Upon arrival at the laboratory, parents and children were invited
to a playroom. Parents were informed about the general procedure and asked not to interfere during the testing phase. The
time spent in the playroom also served to acquaint the child
with the adult experimenter and took no more than 5 min.
Then, the experimenter accompanied the parent and child to
the testing room. In the testing room, the play setting (i.e. the
apparatus with the window-frame) had been arranged on a
small table located in the middle of the room. The parent was
seated at one end of the play table. The child was invited to sit
on the parent’s lap while the experimenter sat down opposite
parent and child behind the plexi-glass window. The parent
was instructed not to interfere during the experimental session
and to minimize interaction with their child as much as possible.
On the experimenter’s side of the play setting, out of reach for the
child, a red stick (height 18 cm) was located directly in front of
and aligned with the centre of the window-frame.
Prior to the participant’s arrival, two keys had been placed
on the child’s side of the play setting, one on the left and one
on the right table edge. Before initiating the joint-action task,
the experimenter reached out with her right hand next to the
window-frame and asked the child to pass her the key on the
child’s left. Analogously, the experimenter repeated the request
on the other side. The purpose of this initial phase was to
ensure that children experienced that they could pass objects
on either side of the window-frame.
Then, the experimenter started the joint-action task by putting on and introducing one of two hand puppets (either a
horse or an elephant). Consequently, one of the experimenter’s
hands was occupied by holding the puppet. The side with
which she started holding the hand puppet was counterbalanced
between children. A plastic cup was fixed to each hand puppet
and the experimenter introduced the hand puppet by saying
‘This is a [horse/elephant] and he has a cup’. Simultaneously to
this introduction, she pointed to the cup. Following this, the
experimenter made the hand puppet balance the cup on top of
the red stick. Then the experimenter said, ‘You know what we
can do with this? We can build a tower together with cups’. To
demonstrate further, the experimenter took a cup from behind
the apparatus, lifted it to make it visible for the child, and stacked
it onto the first cup. She then said that she had no more cups and
asked the child whether he or she had another cup to pass to her
so she could put it on top of the tower for them. At the same
time, the experimenter operated the turning mechanism of the
apparatus such that a cup would appear in front of the child.
While awaiting the child’s response, the experimenter rested
her free arm alongside her body. The experimenter remained
still until the child made a response choice and only then reached
out for the cup.
The response choice of the child (i.e. the side on which they
passed the cup to the experimenter) affected the experimenter’s
subsequent action execution and the smoothness of the joint
action: if the child passed the cup on the experimenter’s
free-hand side, the experimenter could easily reach for it and
subsequently stack it on the tower in the middle (figure 1,
middle). If the child passed the cup on the other side, however,
the experimenter had to bend over with her free hand and stretch
across her occupied arm, with which she balanced the tower on
the stick to receive the cup and stack it on top of the tower
(figure 1, bottom). Note that either response choice of the child
3
Phil. Trans. R. Soc. B 371: 20150371
2. Material and methods
actors sat opposite each other on either side. The actors were separated by a window frame (height 87 cm, width 80 cm) that was
mounted on top of the apparatus and required the actors to
reach around the frame on one of its sides to hand over the
cups (figure 1). All children were engaged in this joint-action
task in a dyadic cooperation with the adult experimenter.
rstb.royalsocietypublishing.org
more challenging and emerge slightly later than adjusting
one’s action plans in the course of a joint action. In particular,
in a study on planning in a collaborative problem-solving
context, Warneken et al. found that 3-year-old children
needed experience with a task before they planned ahead in
a way that integrated the role of their joint-action partner in solving the task. By the age of 5 years, children planned their
actions in this collaborative problem-solving context without
prior experience [28]. Additionally, Gerson et al. demonstrated
that 3-year-old children had more difficulties initially planning
their own actions in a joint context than when they acted on
their own. Their ability to correct their actions to still reach a
common goal, however, was not affected [29]. To what extent
children’s action planning as investigated in these studies
reflects their efforts to reach their own goal rather than planning their actions to accommodate those of another person
remains unclear.
To investigate whether and from which age on children
accommodate their action partner’s actions, we tested children
between the ages of 212 and 5 years in a simple joint-action
game. In this game, children built four towers together with
an adult action partner. For each of the four towers, the
action constraints of their partner changed. The joint-action
game was designed such that children of all age groups
could successfully complete the task. As the task could be
completed successfully in any case, we could exclude the possibility that their behaviour would merely reflect that they were
using the partner to reach the game’s goal. However, how children chose to pass the building elements of the tower to their
partner affected their partner’s action execution.
Downloaded from http://rstb.royalsocietypublishing.org/ on July 28, 2017
joint-action task
4
rstb.royalsocietypublishing.org
Phil. Trans. R. Soc. B 371: 20150371
starting point
response choice:
side of joint-action partner’s free hand
response choice:
side of joint-action partner’s occupied hand
Figure 1. Example of a 3-year-old child engaged in the joint-action task with an adult experimenter. (Top) The start of a trial in which the child has a cup and sees
the experimenter holding the cup tower in her left hand. The lower two images illustrate the different possible response choices of the child. The child chooses in
one trial to pass the cup on the side that is accommodating the other’s actions (middle); and in another trial, on the side that is not accommodating the execution
of the partner’s actions (bottom).
led to a successful goal accomplishment (i.e. stacking the cups to
a tower), but importantly, the child’s action choice differentially
affected the execution of their joint-action partner’s action, either
by accommodating or by impeding it.
After stacking the cup onto the tower, the experimenter
initiated the next trial by asking the child for another cup. To
build one tower, children had to pass five cups. Then the experimenter introduced the next hand puppet with the same
procedure. With each change of hand puppet, she switched
sides by putting the puppet on her previously free hand. Consequently, children now needed to change the side on which they
passed the cup to accommodate their joint-action partner. Which
of the experimenter’s hands were free was balanced over four
blocks. Thus, in total four towers were built, two on each side,
adding up to 20 trials per child. A video recording was made
during the experimental session to code the child’s response
choices offline.
(d) Data processing
Children’s response choices (i.e. the side children chose to pass
the cup to their joint-action partner: the partner’s free or
occupied hand) were coded per trial offline. Figure 1 illustrates
the two possible response choices of the child. Three children
did not complete the last of the four blocks. Moreover, individual
trials were excluded if the child did not successfully pass the cup
to the experimenter (e.g. the cup fell or the parent passed the
cup) or if the child stood up or walked around to pass the cup.
Downloaded from http://rstb.royalsocietypublishing.org/ on July 28, 2017
(a) Initial response choice
To test whether children differed in their initial response
across different ages, we submitted the initial response
choice values to a Pearson’s x2-test. No significant association
between children’s initial response choices and their age was
observed, x23 ¼ 1:048, p ¼ 0.79. Figure 2 illustrates that 64% of
the 212-year-olds, 70% of the 3-year-olds, 77% of the 312-yearolds and 73% of the 5-year-olds passed the cup on the first
trial in such a way that their joint-action partner could
easily reach for the cup with her free hand.
As no age differences were observed, we collapsed data
across age to test whether children were more likely to pass
the cup to their joint-action partner on her free-hand side
on the first trial. Overall, children significantly planned
their actions proactively in a way that accommodated their
partner’s actions, as indicated by the fact that 71% of all
children chose to pass the first cup to their joint-action
partner on her free-hand side ( p , 0.01).
(b) Continuous response choice
The second question we examined was to what extent children of different age groups would flexibly adjust their
action plans over the course of the experiment to
90
percentage of children
80
5
side of joint
action partner’s
available hand
occupied hand
70
60
50
40
30
20
10
0
2½-year-olds 3-year-olds 3½-year-olds 5-year-olds
age group
proportion of mean continuous response
choice (95% CI)
Figure 2. Percentage of children from the four age groups who made the
initial response choice in a way that was accommodating (green) or not
(blue) their joint-action partner’s actions (i.e. reaching and grasping the
cup on the free-hand side).
0.65
**
**
p = 0.05
0.60
p = 0.12
*
p = 0.04
0.55
p = 0.18
0.50
0.45
2½-yearolds
3-year- 3½-yearolds
olds
age group
5-yearolds
Figure 3. Continuous response choices throughout the joint-action task in
proportion (1 representing flexible adjustment to the joint-action partner
in all trials; 0 representing always choosing the opposite response, not comfortable for the joint-action partner). Error bars represent 95% confidence
interval. *p , 0.05, **p , 0.01.
accommodate the actions of their partner. For this purpose,
we tested with a Kruskal– Wallis test and Jonckheere’s
trend test whether children flexibly adjusted their continuous
response and whether this improved with age. The results are
shown in figure 3. Continuous response choice values can
range from 0 (i.e. making the partner’s actions more difficult
on all trials) to 1 (i.e. accommodating the partner’s actions on
all trials). If children continuously passed the cup on one side
it resulted in a continuous response choice value of 0.5.
Children’s continuous response choices differed significantly between age groups, H(3) ¼ 7.9, p , 0.05. Jonckheere’s
test revealed a significant trend in the data: the continuous
response choice values increased with age, J ¼ 1759.5, p ,
0.01, z ¼ 2.62, r ¼ 0.28. To test children’s performance against
chance, we entered the data in a one-sample Wilcoxon
signed-rank test, separately per age group. The 212-year-olds
(M ¼ 0.51, s.d. ¼ 0.03) and 3-year-olds (M ¼ 0.54, s.d. ¼ 0.12)
did not differ from chance (i.e. 0.5) in their averaged continuous
response choices ( ps . 0.05): they equally often passed the cup
to the side of their joint-action partner’s free and occupied
Phil. Trans. R. Soc. B 371: 20150371
3. Results
100
rstb.royalsocietypublishing.org
Six trials were invalid because of an experimenter’s fault
(i.e. counterbalancing error or moving her arm before the children
had finished their reaching movement). Together, due to these
exclusion criteria approximately 2% of all trials were excluded.
We focused our analysis on two measures of interest: the
initial response choice (i.e. the first trial only) and the continuous
response choice (i.e. all trials except the first). In their initial
response choice, children had to plan ahead without having
experienced how their own action would affect their jointaction partner. Thus, the initial response choice reflects whether
children proactively planned to accommodate their joint-action
partner. One child had to be excluded from further analysis
because of walking around the table in the first trial and,
therefore, not providing valid data for the initial response choice.
The continuous response choice, on the other hand, reflects
the extent to which children flexibly adjusted their action plans
to accommodate the other’s action execution over the course of
the joint action. To continuously accommodate the actions of
the joint-action partner, children needed to flexibly switch the
sides on which they passed the cup between blocks. For the continuous response choice, the proportion of trials was calculated
in which children chose the joint-action partner’s free-hand side.
The data for both measures were tested for normality. As the
assumption of normality was violated, non-parametric statistical
tests were used in the subsequent analyses. We tested for main
effects and age differences of each measure separately. We hypothesized that with increasing age, children would adjust their
action plans more to accommodate their joint-action partner. We
expected this development in children’s adjustment of their
action plans to slightly precede their development in proactive
planning. In addition, both measures were entered in one statistical
comparison with test for an interaction of children’s initial and continuous response choices. Both measures assess whether children
take the other into account in their action plans, but address different facets of such planning. We thus hypothesized that children
who initially planned their action according to their partner’s
actions would perform better throughout the subsequent course
of the joint-action game.
Finally, we investigated the relationship between initial and
continuous response choice to determine whether children
who accommodated their action partner from the start would
also show greater adjustments to their partner over the
course of the experiment. To test for this possibility, we computed an independent samples test (Mann–Whitney) with
the initial response choice as independent and the continuous
response choice as dependent variable. Thus, we assigned children who did display accommodating behaviour on the first
trial and those who did not to separate groups. Children who
initially passed the cup to their joint-action partner on her
free-hand side showed a trend towards more accommodating performance as indicated by the continuous response
choice values during the remaining course of the joint action
(M ¼ 0.54, s.d. ¼ 0.09) than children who initially passed
the cup on her occupied-hand side (M ¼ 0.52, s.d. ¼ 0.06,
U ¼ 673.0, z ¼ 21.7, p ¼ 0.08, r ¼ 20.02). Per group, a onesample Wilcoxon signed-rank test was used to assess whether
children’s first response differed from chance. As displayed in
figure 4, children who did not spontaneously hand over the
cup during their first response also did not differ from
chance in how they passed the cup in the entire subsequent
joint-action game ( p ¼ 0.14). On the contrary, if they had
passed the cup initially in a way that accommodated the
execution of their partner’s actions, children were above
chance in their subsequent joint-action performance, passing
the cup more often on the joint-action partner’s free-hand
side ( p , 0.01). Together, these results indicate that children
who initially planned their action according to their partner’s
action execution performed better throughout the subsequent
course of the joint-action game.
4. Discussion
Smooth joint coordination between two actors requires the
planning and adjusting of one’s own actions to those of a
joint-action partner. In the current experiment, we studied
the behaviour of 212- to 5-year-old children in a simple jointaction game to investigate whether and at what age children
plan their own actions to accommodate a joint-action partner.
In particular, we assessed two aspects of children’s action
planning: their proactive planning in the absence of prior
experience with a particular joint action (i.e. initial response
choice); and their flexible adjustment of their action plans to
accommodate their joint-action partner when the constraints
*
**
p = 0.00
0.55
p = 0.14
0.50
0.45
occupied hand
available hand
initial response choice
Figure 4. Continuous response choice throughout the joint-action task separated by children’s initial response choice accommodating (green) or not
(blue) to their joint-action partner initially. Please note that children’s initial
response choice is excluded from their subsequent continuous response
choice. Children who initially chose to pass the cup on the joint-action
partner’s free-hand side tended to show more accommodating behaviour
throughout the joint-action task ( p ¼ 0.08). In addition, children who
started out passing the cup on the joint-action partner’s occupied side did
not significantly adjust their behaviour throughout the task to accommodate their joint-action partner. Error bars represent 95% confidence
interval. *p , 0.05, **p , 0.01.
of their partner’s but not their own actions changed throughout the joint play (i.e. continuous response choice, measured
over the course of the experiment).
This study yielded two major findings. First, we found evidence that young children proactively planned their actions
such that they accommodated the actions of their joint-action
partner. Second, our findings suggest developmental increase
in children’s ability to flexibly adjust their action plans to
accommodate their joint-action partner. Moreover, 212- and 3year-old children did not differ from chance in the way they
planned their actions with respect to their joint-action partner
across the joint-action game. At the age of 312 years, children
showed flexible adjustment to the partner’s constraints as
their performance differed from chance. By 5 years, children
were clearly above chance levels, reflecting that they flexibly
accommodated to their action partner. However, even
5-year-olds only did so in less than 60% of the trials.
(a) Accommodating the other’s actions: children’s
proactive action planning
At the very start of the joint-action game, 70% of all children
passed the cup to the joint-action partner on her free-hand
side. Doing so assured that the adult partner could easily
reach for and subsequently stack the cup onto the tower. The
lack of developmental differences in this regard was somewhat
unexpected as we had hypothesized an age-related increase in
children’s proactive planning and a potentially delayed onset
of proactive planning skills compared with flexible adjustment
of their action plans. It also contradicts some of the previous
developmental research [12,13,28]. For instance, Warneken
et al. [28] found that 3-year-olds fail to plan their actions
ahead in relation to another person’s actions in a collaborative
problem-solving task. Based on this research, we had expected
that especially the younger children in our study would have
greater difficulties planning their actions to accommodate
Phil. Trans. R. Soc. B 371: 20150371
(c) Linking initial and continuous response choice
6
0.60
rstb.royalsocietypublishing.org
hands. Whereas the 3-year-old children show more choices
towards the free-hand side of their partner, their variability
in performance was relatively high. By the age of 312 years,
children did pass the cup significantly more often to their
joint-action partner’s free-hand than to the occupied-hand
side (M ¼ 0.53, s.d. ¼ 0.07, p ¼ 0.04). By the age of 5, children
then showed a further increase in their tendency to accommodate their action partner flexibly (M ¼ 0.56, s.d. ¼ 0.1,
p ¼ 0.005). The findings thus indicate that with increase in
age, children adjusted their action plans more flexibly to
accommodate the other’s action execution. Yet although
5-year-old children significantly adjusted their action plans,
the extent to which they did so is still very limited. On average, they handed the cup over on the side their partner
could easily retrieve it only in 56% of all trials.
proportion of mean continuous response
choice (95% CI)
Downloaded from http://rstb.royalsocietypublishing.org/ on July 28, 2017
Downloaded from http://rstb.royalsocietypublishing.org/ on July 28, 2017
In contrast with children’s proactive action planning, the
flexible adjustment of their action plans improved during
early childhood. Children aged 212 years hardly ever adjusted
their actions but instead consistently passed the cup to their
joint-action partner on the same side they had chosen
initially. They did so irrespective of its disadvantageous
effect on their partner’s action execution. Three-year-old
children did not differ from chance performance either, but
showed a descriptively higher proportion of trials in which
they adjusted their actions to accommodate their partner.
The high variability in this age group might indicate that
they are in a transition towards a more stable and consistent
accommodation behaviour. By the age of 312 years children
performed significantly above chance in terms of flexibly
integrating their joint-action partner’s constraints in their
action plans. By the age of 5 years children clearly showed
flexible adjustment of their action plans during the joint
action. More surprising, however, was the generally low
extent with which children adjusted their action plans to
accommodate their partner’s actions. Even by the age of 5,
7
Phil. Trans. R. Soc. B 371: 20150371
(b) Accommodating the other’s actions: children’s
flexible adjustment of their action plans
children adjusted their plans to the other person in less than
60% of the trials. In other words, children adjusted their
response on average only on 2 of the 10 trials that differed
from their initial response. It therefore remains an open question when children reach adult levels in flexible adjustment
of their action plans to accommodate others. It is important
to mention, however, that we found in a previous study on
joint-action planning that adults adjust their actions less
consistently to their action partner if the joint action requires
higher-order planning [3]. At first glance, our findings seem
at odds with previous findings demonstrating prosocial behaviour already in infancy and early childhood (e.g. [40,41] for a
review). For instance, at the age of 18 months, toddlers open
a door for another person or hand objects that are out of
reach for the other. Most of these studies, however, assess
whether infants and young children engage in an interaction
to assist another person in reaching his or her goal, rather
than examining how children adjust their action plans for
another person when already engaged in a joint action (as
was the case in the current study). Flexibly adjusting action
plans for a joint-action partner during an ongoing joint action
might be far more demanding and thus develop only during
and beyond early childhood. A combination of higher cognitive abilities, such as inhibitory control, cognitive flexibility
and perspective taking, might be needed to adjust one’s
plans flexibly to accommodate the joint-action partner. While
children’s early proactive planning to accommodate others’
actions is in line with findings of early prosocial behaviour,
the dynamically changing constraints for the other person
might put more demands on their higher cognitive abilities
to achieve flexible adjustment in action plans.
The ongoing development of executive function skills, such
as inhibitory control and cognitive flexibility, probably plays an
important role in the developmental pattern we found. Notably, the current joint-action game was designed with inherent
inhibition and flexibility demands which are typical but not
necessary for joint-action tasks. The change between blocks
required children to refrain from choosing the same side
and to switch to the other side to account for where their
joint-action partner’s free hand was. The ability to inhibit a
prepotent response and to flexibly adjust a previously established response pattern underlies top-down processes
associated with frontal brain areas. These areas, which are still
in maturation up to adolescence, are thought to be involved
in inhibitory control and cognitive flexibility [42–45]. Improvement in the performance of tasks requiring inhibitory control
(such as the gift delay task) and cognitive flexibility (such as
the Stroop or the dimensional change card-sorting task) is consistently found between the ages of 3 and 5 years [26,45,46]. It is
around the same ages that we observed flexible adjustment
emerging in children’s action plans with respect to the changing
constraints of their joint-action partner. Although far from conclusive, the increased variability of the 3-year-old children’s
behaviour is consistent with this developmental change. As a
group, however, they still performed at chance level.
As discussed in the context of proactive action planning,
we assume that in addition to and possibly intertwined with
inhibitory control and cognitive flexibility, perspective-taking
skills might play an important role in adjusting one’s action
plans to another person. This holds true especially for
the flexible adjustment needed to account for the other’s
changing task constraints. In order to plan their own actions
to accommodate the actions of their partner, children might
rstb.royalsocietypublishing.org
others. Our findings, however, suggest that young children
detected the current action affordances of the other person
and planned their actions accordingly from the start of the
experiment. In other words, children perceived that the grasping and stacking of the cup had to be done with the partner’s
free hand.
One possibility is that children used visual spatial cues
from the initial demonstration of the action. More specifically,
although not having experienced how their own actions affect
the other person in this task, children might have based their
detection of the other’s affordances on their observations of
how the other person performed the action herself. Recent findings based on kinematic recordings suggest that adults adjust
their own movements according to the action affordances present in social contexts [30]. For example, the kinematics of
the hand and body postures of another person in relation to
objects provide cues about action affordances for adult observers [31–34]. In our task, children’s perception of affordances
and consequently their proactive planning may also reflect
the use of visual spatial cues about their joint-action partner.
Furthermore, in the current study children sat on the lap of
their parent, who may have subconsciously biased their
children with their own body movements. Although parents
were instructed not to interfere in any way during the testing,
potential subconscious movements of the parents might have
contributed to children’s behaviour.
Another possibility is that children’s perception of the
other’s affordances is based on early forms of perspective
taking. Previous findings show that by the age of 3 years,
children can master visual perspective taking [35]. Development of visual perspective taking precedes the emergence of
epistemic perspective taking (as part of the theory of mind
development), which typically develops during and beyond
early childhood [36–39]. In the current game, children might
have used a simple form of visual perspective taking to perceive the action affordances of their partner. Follow-up
research is needed to investigate the potential contribution of
perspective-taking skills on the detection of action affordances
in children’s proactive planning to accommodate others.
Downloaded from http://rstb.royalsocietypublishing.org/ on July 28, 2017
We also found a marginally significant effect indicating a
potential relationship between the two measures of proactive
planning and flexible adjustment of their action plans: the
children who at the very beginning of the joint action proactively planned their actions in a way that accommodated the
actions of their partner tended to be more flexible in adjusting
their actions over the course of the experiment. Although the
two measures reflect distinct aspects of how children plan
their actions to accommodate their partner’s actions, they
might build upon each other.
Potentially, children’s detection of the other’s action
affordances forms the basis for their initial planning and a
prerequisite of their adjustment to accommodate their
partner. On top of detecting the other’s action affordances,
however, ongoing dynamic changes during the joint action
put additional demands on higher cognitive abilities, such as
inhibitory control and cognitive flexibility, which leads to the
delayed development of children’s flexible adjustment. This
might explain why children who detect the affordances of the
other’s actions initially plan their actions accordingly and also
show a higher performance in flexibly adjusting their plans.
5. Summary
Together, the current findings indicate that children proactively
plan their actions in a way that accommodates the actions of
Ethics. All parents provided written informed consent for participation of their child in this study. The research was approved by
the local ethics committee (http://www.ru.nl/socialsciences/ethicscommittee/ethics-committee; ECG 2012-1301-006).
Data accessibility. The datasets supporting this article have been
uploaded as part of the electronic supplementary material.
Authors’ contributions. M.M. contributed to conception and design,
acquisition of data, analysis and interpretation of data, drafting the
article, and revising it critically for important intellectual content.
R.P.R.D.v.d.W. contributed to conception and design, interpretation
of data, and drafting and revising the article. S.H. contributed to
conception and design, interpretation of data, and drafting and revising the article. All authors approved the final version of the
manuscript.
Competing interests. We have no competing interests.
Funding. We received no funding for this study.
Acknowledgements. We thank all the children and parents for participating in this study. Moreover, we thank our laboratory manager,
Angela Khadar, for her help during the study. We also appreciate
the help during data acquisition of Celeste Peereboom. We also
acknowledge the technical support in building the apparatus for
this study provided by the Technical Support Group at Radboud
University, Nijmegen.
References
1.
2.
3.
4.
5.
Ray M, Welsh TN. 2011 Response selection during a
joint action task. J. Mot. Behav. 43, 329–332.
(doi:10.1080/00222895.2011.592871)
Gonzalez DA, Studenka BE, Glazebrook CM, Lyons
JL. 2011 Extending end-state comfort effect: do we
consider the beginning state comfort of another?
Acta Psychol. 136, 347–353. (doi:10.1016/j.actpsy.
2010.12.009)
Meyer M, van der Wel RP, Hunnius S. 2013 Higherorder action planning for individual and joint object
manipulations. Exp. Brain Res. 225, 579– 588.
(doi:10.1007/s00221-012-3398-8)
Cirelli LK, Einarson KM, Trainor LJ. 2014
Interpersonal synchrony increases prosocial behavior
in infants. Dev. Sci. 17, 1003 –1011. (doi:10.1111/
desc.12193)
Tunçgenç B, Cohen E, Fawcett C. 2015 Rock with
me: the role of movement synchrony in infants’
social and nonsocial choices. Child. Dev. 86,
976 –984. (doi:10.1111/cdev.12354)
6. Marsh KL, Richardson MJ, Schmidt RC. 2009 Social
connection through joint action and interpersonal
coordination. Top. Cogn. Sci. 1, 320– 339. (doi:10.
1111/j.1756-8765.2009.01022.x)
7. Hove MJ, Risen JL. 2009 It’s all in the timing:
interpersonal synchrony increases affiliation. Soc.
Cogn. 27, 949– 960. (doi:10.1521/soco.2009.27.6.
949)
8. Wiltermuth SS, Heath C. 2009 Synchrony and
cooperation. Psychol. Sci. 20, 1 –5. (doi:10.1111/j.
1467-9280.2008.02253.x)
9. Brownell CA. 2011 Early developments in joint
action. Rev. Phil. Psychol. 2, 193 –211. (doi:10.
1007/s13164-011-0056-1)
10. Bigelow AE, Best C. 2012 Peek-a-what?
Infants’ response to the still-face task after
11.
12.
13.
14.
normal and interrupted peek-a-boo. Infancy
18, 400 –413. (doi:10.1111/j.1532-7078.2012.
00124.x)
Warneken F, Chen F, Tomasello M. 2006 Cooperative
activities in young children and chimpanzees. Child
Dev. 77, 640–663. (doi:10.1111/j.1467-8624.2006.
00895.x)
Gräfenhain M, Behne T, Carpenter M, Tomasello M.
2009 Young children’s understanding of joint
commitments. Dev. Psychol. 45, 1430–1443.
(doi:10.1037/a0016122)
Hamann K, Warneken F, Tomasello M. 2012
Children’s developing commitments to joint goals.
Child Dev. 83, 137–145. (doi:10.1111/j.1467-8624.
2011.01695.x)
Aureli T, Presaghi F. 2010 Developmental trajectories
for mother-infant coregulation in the second year of
life. Infancy 15, 557 –585.
8
Phil. Trans. R. Soc. B 371: 20150371
(c) Relationship between children’s proactive planning
and their flexible adjustment of action plans
their partner early in childhood. By contrast, the flexible adjustment of their action plans to their partner only begins to develop
in the fourth and fifth year of life. Notably, even at the age of
5 children only adjusted their action plans to a surprisingly
small degree. To our knowledge, this is the first study on
when children begin to plan and adjust their actions beyond
the purpose of reaching a goal to accommodate their partner’s
actions. As this is the first study on the development of planning
own actions to accommodate others, complementary findings
from additional experimental settings are required to draw
general conclusions about the developmental trajectory. Underlying processes, such as inhibitory control, cognitive flexibility
and perspective taking, may play a crucial role in learning
how to plan one’s own actions in a way that accommodates
the actions of a joint-action partner and brings about smooth
joint-action coordination. Further research is needed to examine
the precise contributions of each of these processes.
rstb.royalsocietypublishing.org
need to take the other’s perspective and track potential changes
in the constraints of the other’s actions (like holding the tower
on different sides). The observation that children in our study
flexibly adjusted their actions to the action partner to an
increasing degree parallels children’s emerging epistemic
perspective-taking skills. Still, future research is required to
test direct links between the development of perspectivetaking and children’s action planning to accommodate their
joint-action partner.
Downloaded from http://rstb.royalsocietypublishing.org/ on July 28, 2017
36. Flavell JH, Flavell ER, Green FL, Moses LJ. 1990
Young children’s understanding of fact beliefs
versus value beliefs. Child Dev. 61, 915– 928.
(doi:10.1111/j.1467-8624.1990.tb02831.x)
37. Wimmer H, Pemer J. 1983 Beliefs about beliefs:
representation and constraining function of wrong
beliefs in young children’s understanding of
deception. Cognition 13, 103–128. (doi:10.1016/
0010-0277(83)90004-5)
38. Wellman HM, Cross D, Watson J. 2001 Meta-analysis
of theory-of-mind development: the truth about
false belief. Child Dev. 72, 655–684. (doi:10.1111/
1467-8624.00304)
39. Blakemore SJ. 2008 The social brain in adolescence.
Nat Rev Neurosci. 9, 267–277. (doi:10.1038/nrn2353)
40. Paulus M. 2014 The emergence of prosocial
behavior: why do infants and toddlers help,
comfort, and share? Child Dev. Perspect. 8, 77 –81.
(doi:10.1111/cdep.12066)
41. Zelazo PD, Carlson SM. 2012 Hot and cool executive
function in childhood and adolescence:
development and plasticity. Child Dev. Perspect. 6,
354–360. (doi:10.1111/j.1750-8606.2012.00246.x)
42. Casey BJ et al. 1997 A developmental functional
MRI study of prefrontal activation during
performance of a go-no-go task. J Cognitive
Neurosci. 9, 835–847. (doi:10.1162/jocn.1997.
9.6.835)
43. Casey BJ et al. 2011 Behavioral and neural correlates
of delay of gratification 40 years later. Proc. Natl
Acad. Sci. USA 108, 14 998– 15 003. (doi:10.1073/
pnas.1108561108)
44. Durston S, Thomas KM, Yang YH, Ulug AM,
Zimmerman RD, Casey BJ. 2002 A neural basis for
the development of inhibitory control. Dev. Sci. 5,
F9 –F16. (doi:10.1111/1467-7687.00235)
45. van Bers BMCW, Visser I, van Schijndel TJP, Mandell
DJ, Raijmakers MEJ. 2011 The dynamics of
development on the dimensional change card
sorting task. Dev. Sci. 14, 960–971.
46. Jacques S, Zelazo PD. 2005 On the possible roots of
cognitive flexibility. In The development of social
cognition and communication (eds BD Homer,
CS Tamis-LeMonda), pp. 53 –81. Mahwah, NJ:
Lawrence Erlbaum Associates Publishers.
9
Phil. Trans. R. Soc. B 371: 20150371
26. Carlson SA. 2005 Developmentally sensitive
measures of executive function in preschool
children. Dev. Neuropsychol. 28, 595–616. (doi:10.
1207/s15326942dn2802_3)
27. Kochanska G, Murray KT, Harlan ET. 2000 Effortful
control in early childhood: continuity and change,
antecedents, and implications for social
development. Dev. Psychol. 36, 220–232.
28. Warneken F, Steinwender J, Hamann K, Tomasello
M. 2014 Young children’s planning in a
collaborative problem-solving task. Cogn. Dev. 31,
48 –58. (doi:10.1016/j.cogdev.2014.02.003)
29. Gerson S, Hunnius S, Bekkering H. 2013 Playing for
us: the influence of joint action on planning in
three-year-olds. In Proc. of the 35th Annual Conf. of
the Cognitive Science Society (eds M Knauff,
M Pauen, N Sebanz, I Wachsmuth), pp. 2380–
2385. Austin, TX: Cognitive Science Society.
30. Scorolli C, Miatton M, Wheaton LA, Borghi AM.
2014 I give you a cup, I get a cup: a kinematic
study on social intention. Neuropsychologia 57,
196 –204. (doi:10.1016/j.neuropsychologia.2014.03.
006)
31. Ferri F, Stoianov I, Gianelli C, D’Amico L, Borghi AM,
Gallese V. 2010 When action meets emotions. How
facial displays of emotion influence goal-related
behavior. PLoS ONE 5, e13126. (doi:10.1371/journal.
pone.0013126)
32. Becchio C, Sartori L, Bulgheroni M, Castiello U. 2008
Both your intention and mine are reflected in the
kinematics of my reach-to-grasp movement.
Cognition 106, 894–912. (doi:10.1016/j.cognition.
2007.05.004)
33. Sartori L, Becchio C, Castiello U. 2011 Cues to
intention: the role of movement information.
Cognition 119, 242–252. (doi:10.1016/j.cognition.
2011.01.014)
34. Sartori L, Straulino E, Castiello U, Avenanti A. 2011
How objects are grasped: the interplay between
affordances and end-goals. PLoS ONE 6, e25203.
(doi:10.1371/journal.pone.0025203)
35. Moll H, Meltzoff AN. 2011 How does it look? Level
2 perspective-taking at 36 months of age. Child Dev.
82, 661– 673. (doi:10.1111/j.1467-8624.2010.
01571.x)
rstb.royalsocietypublishing.org
15. Vesper C, Butterfill S, Knoblich G, Sebanz N. 2010
A minimal architecture for joint action. Neural Netw.
23, 998–1003. (doi:10.1016/j.neunet.2010.06.002)
16. Milward SJ, Kita S, Apperly IA. 2014 The development
of co-representation effects in a joint task: do children
represent a co-actor? Cognition 132, 269–279.
(doi:10.1016/j.cognition.2014.04.008)
17. Saby JN, Bouquet CA, Marshall PJ. 2014 Young
children co-represent a partner’s task: evidence for a
joint Simon effect in five-year-olds. Cogn. Dev. 32,
38 –45. (doi:10.1016/j.cogdev.2014.08.001)
18. Warneken F, Gräfenhain M, Tomasello M. 2012
Collaborative partner or social tool? New evidence
for young children’s understanding of joint
intentions in collaborative activities. Dev. Sci. 15,
54 –61. (doi:10.1111/j.1467-7687.2011.01107.x)
19. Elsner C, D’Ausilio A, Gredebäck G, Falck-Ytter T,
Fadiga L. 2013 The motor cortex is causally related
to predictive eye movements during action
observation. Neuropsychologia 51, 488 –492.
(doi:10.1016/j.neuropsychologia.2012.12.007)
20. Schubotz RI, von Cramon DY. 2002 Predicting
perceptual events activates corresponding motor
schemes in lateral premotor cortex: an fMRI
study. NeuroImage 15, 787 –796. (doi:10.1006/
nimg.2001.1043)
21. Frith CD, Frith U. 2006 The neural basis of
mentalizing. Neuron 50, 531– 534. (doi:10.1016/j.
neuron.2006.05.001)
22. Cross ES, Hamilton AFDC, Kraemer DJ, Kelley WM,
Grafton ST. 2009 Dissociable substrates for body
motion and physical experience in the human action
observation network. Eur. J. Neurosci. 30, 1383–
1392. (doi:10.1111/j.1460-9568.2009.06941.x)
23. Sohn M-H, Ursu S, Anderson JR, Stenger VA, Carter CS.
2000 The role of prefrontal cortex and posterior parietal
cortex in task switching. Proc. Natl Acad. Sci. USA 97, 13
448–13 453. (doi:10.1073/pnas.240460497)
24. Meyer M, Bekkering H, Paulus M, Hunnius S. 2010
Joint action coordination in 212- and 3-year-old
children. Front. Hum. Neurosci. 4. (doi:10.3389/
fnhum.2010.00220)
25. McCormack T, Atance CM. 2011 Planning in young
children: a review and synthesis. Dev. Rev. 31,
1– 31. (doi:10.1016/j.dr.2011.02.002)