Animal Behaviour 86 (2013) 755e761
Contents lists available at ScienceDirect
Animal Behaviour
journal homepage: www.elsevier.com/locate/anbehav
Chimpanzees, Pan troglodytes, recognize successful actions, but fail
to imitate them
David Buttelmann a, b, *, Malinda Carpenter b, Josep Call b, Michael Tomasello b
a
b
Research Group ‘Kleinkindforschung in Thüringen’, University of Erfurt, Germany
Department of Developmental and Comparative Psychology, Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany
a r t i c l e i n f o
Article history:
Received 15 October 2012
Initial acceptance 19 December 2012
Final acceptance 9 July 2013
Available online 16 August 2013
MS. number: 12-00792R
Keywords:
attention
chimpanzee
intention understanding
nonhuman primate
Pan troglodytes
social learning
Cultural transmission, by definition, involves some form of social learning. Chimpanzees and other
nonhuman primates clearly engage in some forms of social learning enabling some types of cultural
transmission, but there is controversy about whether they copy the actual bodily actions of demonstrators. In this study chimpanzees recognized when a human actor was using particular bodily actions
that had led to successful problem solving in the past. But then when it was their turn to solve the
problem, they did not reproduce the human actor’s bodily actions themselves, even though they were
clearly capable of producing the movements. These results help us identify more precisely key reasons
for the differences in the social learning and cultural transmission of humans and other primates.
Ó 2013 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
Learning by observing others interacting with the environment
enables organisms to acquire important information without
engaging in the slower and less efficient process of trial-and-error
learning (Laland 2004; Boyd et al. 2011). There are many different
forms of social learning, mainly differing in terms of exactly what is
learned or replicated (e.g. the action, the result of the action, etc.),
practised by many different species, from fish to birds to primates
(see Galef & Laland 2005). Much of human cultural behaviour is
socially acquired by imitation, that is, the copying of the precise
bodily movements of others (Tomasello 1999; see Caldwell et al.
2012 for an alternative view). Although some researchers
consider human and chimpanzee cultures to be homologous
(Whiten et al. 2003, 2005; Horner et al. 2006), it is unclear whether
action copying plays an important role in chimpanzee social
learning. Indeed, most of the examples of chimpanzee cultural
behaviour involve some form of tool use, whose acquisition may be
based on copying the effect of the tool on the environment
(emulation learning) rather than the actual bodily actions of the
tool user (Tomasello 1996). Even studies suggesting that chimpanzees copy bodily actions (Whiten et al. 2005; Hopper et al.
* Correspondence: D. Buttelmann, Research Group ‘Kleinkindforschung in
Thüringen’, University of Erfurt, Nordhaeuser Strasse 63, D-99089 Erfurt, Germany.
E-mail address: [email protected] (D. Buttelmann).
2008) have different interpretations (Tennie et al. 2009), and
have not been replicated in different chimpanzee groups (Hopper
et al. 2007) or in different tasks (Tennie et al. 2006). When chimpanzees and human children are directly compared in tool use
situations, children copy bodily actions much more often than
chimpanzees (Nagell et al. 1993; Horner & Whiten 2005; Dean et al.
2012). Furthermore, in more social instrumental situations, that is,
communicative situations in which bodily gestures are used for
social ends, both observational and experimental research has
found little evidence of chimpanzees acquiring new gestures by
copying them (Tomasello et al. 1997; Tennie et al. 2012), although
there is positive evidence for this in even very young human children (e.g. Williamson et al. 2013).
However, after extensive training chimpanzees can learn to copy
some bodily actions (Hayes & Hayes 1952; Custance et al. 1995).
Moreover, some chimpanzees recognize when a human is copying
their actions as well (Nielsen et al. 2005; Haun & Call 2008; for
imitation recognition in monkeys see Paukner et al. 2005, 2009). So
the question arises why chimpanzees do not copy the actions of
others more readily in instrumental situations such as tool use and
communicative gestures. Perhaps in instrumental (problem-solving) situations, they simply do not attend at all to the actions of
others (they need training or special situations such as someone
copying them to attend to actions). While many animals can pay
attention to even very subtle behavioural cues (see e.g. Pfungst
0003-3472/$38.00 Ó 2013 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.anbehav.2013.07.015
756
D. Buttelmann et al. / Animal Behaviour 86 (2013) 755e761
1911), in some circumstances chimpanzees apparently find it very
difficult to use information from human actions to obtain rewards
(e.g. Tomasello et al. 1997; Itakura et al. 1999), so this is a real
possibility. Alternatively, they may attend to the actions, but cannot
or do not translate them into their own actions. Previous research
cannot determine which of these hypotheses might be correct
because the studies in which chimpanzees were trained to mimic
bodily actions involved different individuals and different actions
from the studies in which they failed to copy actions in instrumental contexts.
To investigate these hypotheses directly, we designed the current study with three aims in mind. First, we investigated whether
chimpanzees attend to the actions of humans closely enough to
recognize these actions and respond appropriately to them later
(recognition test). Second, we investigated whether, when subsequently given the opportunity to solve the same problem themselves, individuals who attended to the actions would then use that
information to imitate the actions (imitation test). Finally, we
additionally investigated whether prior self-experience on the task
would increase individuals’ ability to recognize the human’s actions
in the first place (self-experience condition).
METHODS
Procedure
The procedure consisted of an information phase, in one of two
conditions, and two test phases.
Information phase
The information phase consisted of four consecutive sessions of
12 trials per day (for a total of 48 trials). During this phase, chimpanzees were provided with information about one of two techniques (‘sliding’ or ‘lifting’) for opening the apparatuses to extract
the food inside. This information was provided in two different
ways: it was demonstrated by E1 in one condition, and subjects
learned it through direct experience in the other condition. That is,
after baiting the boxes out of the chimpanzees’ view and setting up
the test equipment, E1 stood behind one of the two boxes in front of
the Plexiglas window, called the chimpanzees to attract their
attention and released the first reward into the container, so that it
was visible through the Plexiglas door. The subsequent procedure
differed according to the experimental condition (with subjects
randomly assigned to one of these two experimental conditions,
with age and sex matched as well as possible). Ten subjects
participated in the modelled condition and eight participated in the
self-experience condition (10 were assigned to the latter condition
but two dropped out; see Participants section).
Participants
Eighteen chimpanzees (mean age ¼ 18.1 years, range 3.8e33.0
years, 12 females, six males) participated in this study. Additionally,
two female chimpanzees began the information phase (see below)
but were not tested because they were unwilling to approach the
apparatuses. The apes were housed socially in two groups of at least
six individuals at the Wolfgang Köhler Primate Research Center in
the Leipzig Zoo, Germany. Each group had access to an indoor area
(230 and 430 m2) and an outdoor area (1680 and 4000 m2) furnished with various climbing structures, shelter and natural vegetation. At night, the apes sleep in several series of cages (40 and
50 m2). In addition to experiments, the animals are provided with a
special enrichment programme, including various kinds of tools
and foraging containers. Several times per day, the chimpanzees are
fed a diet consisting primarily of vegetables, fruits and cereals with
regular additions of eggs and meat. Test sessions took place in a
familiar enclosure (approximately 30 m2). The subjects were used
to being separated in adjacent enclosures from their group members for testing. They were not food deprived for testing, and water
was available throughout all test periods. They were not distressed
and were free to stop participating at any time. German law on
animal rights and ASAB/ABS guidelines (ASAB/ABS 2012) were
followed throughout.
Apparatus
Test materials consisted of two plastic tables (80 40 cm),
placed in front of Plexiglas windows opposite each other with a
distance of 20 cm between them. There were two identical apparatuses (modelled after Horner et al. 2006): opaque boxes
(43 40 cm and 16.5 cm high) with a transparent Plexiglas door
(10 10.5 cm) with a small knob (4.5 2 cm) on the front (see
Fig. 1a). This door could be opened in two different ways: it could
either be slid to the (right) side (see Fig. 1b) or lifted (see Fig. 1c) to
give access to a small container inside the box. Within each box
there was a circular reward holder on top of this container (not
visible through the Plexiglas door) that could hold up to six rewards
(grapes used throughout the experiment), by means of which the
experimenter could drop rewards into the container to bait the box
(see Fig. 1d).
Modelled condition. In this condition, the apparatuses stood on the
tables facing the Plexiglas windows, approximately 20 cm from the
windows (see Fig. 2a). E1 demonstrated to the subjects how to use
one of the actions to open the door successfully (see Fig. 2a). After he
ensured that subjects were paying attention (i.e. were sitting behind
the Plexiglas window, facing the apparatus), he slowly opened the
door of the first box by either lifting or sliding (depending on which
action had been assigned to the subject). For lifting, he bent over the
apparatus, grasped the knob on the door with his right hand, and
lifted it as high as possible (90 ), then took the reward with his left
hand, slowly closed the door, and handed the reward to the subject
through a hole at the bottom centre of the Plexiglas window. For
sliding, he bent over the apparatus, grasped the knob with his left
hand, slid the door open as far as possible (8.5 cm), took the reward
with his right hand, closed the door slowly, and handed the reward
to the subject through the same hole as in the other action. After a
reward was handed over, he released the next reward into the
container. Each trial lasted about 5 s with an additional 10 s delay
between the trials. Four of the chimpanzees in this condition saw the
human open the door of the box by lifting it and six saw the human
open the door of the box by sliding it (the numbers are not equal
because two more chimpanzees were planned to observe the lifting
action but they declined to participate in the information phase).
Self-experience condition. In this condition both apparatuses were
attached to the Plexiglas windows. Cut into the windows was a
rectangular hole, such that the knob of the doors protruded into the
chimpanzees’ enclosure, and the doors could easily be opened by
the subjects (see Fig. 2c). During the information phase in this
condition, subjects did not witness a human acting on the apparatuses but instead were given experience in opening the apparatus
themselves, for the same amount of time as the chimpanzees in the
modelled condition had observed the successful human. For each
individual the box was fixed so it could open in only one way. Five
of the chimpanzees in this condition opened the apparatus by
lifting and three opened it by sliding. Once a reward was obtained
from the apparatus, E1 waited approximately 10 s, and then
released the next reward into the container. Already in their first
trial, all subjects managed to obtain the reward very quickly
(means: 4 s for lift and 6 s for slide).
D. Buttelmann et al. / Animal Behaviour 86 (2013) 755e761
757
Figure 1. A depiction of the apparatus used in the study and how it could be opened. (a) The closed box as seen from the subjects’ perspective. (b) The door slid open. (c) The door
lifted open. (d) The apparatus as seen from the experimenter’s perspective.
In both conditions, after subjects had received six trials on one
apparatus, the experimenter repeated this procedure (with the
same action) with the other apparatus for the next six trials, for a
total of 12 trials per day.
Recognition test
For the recognition test, in both conditions, all subjects were
then presented with two new humans, standing next to each other
adjacent to their cage (see Fig. 2b), each attempting unsuccessfully
to open the box using a different technique (i.e. one trying to slide
the door, the other one trying to lift, see below). The recognition
test consisted of a single trial and took place at the end of the 4th
day. Critically, for this test the doors of the apparatuses were locked
shut so that they could not be moved at all; the only information
chimpanzees received was about the experimenters’ bodily
movements. After setting up, the two new experimenters (E2 and
E3) entered the room, stood behind the two apparatuses, bent over
and waited for E1 to signal them to start acting on the apparatuses
(see Fig. 2b). When E2 and E3 were ready, the subjects were let into
the test room, and E1 centred them at the starting position with
fruit juice or food pellets. As soon as subjects were centred, E1 left
the room, and signalled to E2 and E3 to start acting on the boxes.
Each experimenter then grasped the knob of their box and tried
unsuccessfully to open the door, each using a different action:
whereas one tried to lift the door, the other tried to slide it open,
both acting exactly as described for E1 in the modelled condition.
They did so for 120 s. During this time the chimpanzees’ behaviour
was coded (see Coding section). While acting, both experimenters
showed signs of effort (stressed facial expressions and effort
grunts), and occasionally slipped their fingers off the knobs. To keep
the experimenters’ actions as synchronized as possible they were
instructed to mimic each other: after one experimenter slipped his
fingers off or grunted, the other experimenter waited 1 s, and then
acted in the same way, with the initializing role switching
randomly. To control for potential experimenter biases E2 and E3
were blind to the hypotheses of the study and the action used in the
information phase. Additionally, E1 monitored all aspects of their
actions on a small monitor from outside the test room. After 120 s
the test was over. E1 signalled to E2 and E3, who stopped acting and
left the room. Then E1 entered and gave subjects a grape at the
starting position.
The rationale behind this test was that if subjects recognized the
action they had observed to be successful in the past they should
spend more time in close proximity to this action in order to receive
the reward from out of the apparatus. Furthermore, we expected
chimpanzees to spend more time looking at the appropriate action
(the one that had been successful in the information phase). This
prediction might be seen as at odds with predictions made in
violation-of-expectation experiments (e.g. see LaBarbera et al.
1976; Munakata et al. 2001). In experiments of that kind, subjects
tend to look less at familiar stimuli. However, there is an important
difference between typical violation-of-expectation experiments
and the current study. In the current study, in both conditions
subjects had received a food reward from the apparatus in each trial
758
D. Buttelmann et al. / Animal Behaviour 86 (2013) 755e761
Figure 2. The general set-up used in the study. (a) In the modelled condition, an experimenter attracts the subject’s attention before modelling the successful action. (b) The set-up
in the recognition test was identical in both conditions. The chimpanzee sits at the centred starting position, with E2 and E3 standing behind the apparatuses, waiting for E1’s signal
to start the manipulation. (c) In the self-experience condition, a subject gains experience by acting on the apparatus itself.
of the information phase. Thus during the recognition test they
should stay closer to the apparatus that was most likely to result in
food for them, and, since it was a common experience for them that
human experimenters gave them food, they should pay attention to
the person who would be most likely to give them that reward.
Imitation test
After a delay (on average 33 days; range 8e117 days owing to
constraints regarding scheduling of test times), subjects in the
modelled condition (only) were additionally presented with an
imitation test. For this, they first received another session of 12
information phase trials by the original experimenter, in exactly the
same manner as described above. Subjects were taken into an
adjacent occluded enclosure briefly while both apparatuses were
attached to the Plexiglas windows and baited. Then subjects were
let into the test room to act on the apparatuses themselves for the
first time, for 12 consecutive trials (six on each apparatus). As in the
information phase for the self-experience condition, after subjects
retrieved a reward, E1 waited approximately 10 s to release the
next reward. Importantly, for the imitation test the doors of the
apparatuses were not blocked in any way, such that for each trial
both actions, lift and slide, could be used to obtain the rewards from
the containers. The imitation test was not presented to subjects in
the self-experience condition because, in contrast to human
children, chimpanzees tend to ignore social information once they
have relevant personal experience with an apparatus (Kendal et al.
2005; Wood et al. 2013).
The apparatus the experimenter started with during the information phase, the sides of the experimenters at test and the side of
the lifting action at test were counterbalanced, with subjects’
gender and age matched as closely as possible in all cases. Testing
was videotaped.
Coding, Reliability and Analysis
Coding was done by the first author using videotapes such that
condition and action used during the information phase could not
be determined. The corners of large squares (1 m2) had been
marked with paint on the floor inside the enclosure in front of the
two Plexiglas windows the experimenters stood behind in the
recognition test phase (see Fig. 2b). We coded subjects’ proximity
and attention during this test phase, that is, during the 120 s that
the two experimenters acted on the apparatuses. First, proximity
was the accumulated time (s) a subject spent in each square during
the response period, or in other words the amount of time chimpanzees remained in front of each of the two experimenters while
they were trying to open the apparatuses using different actions.
Second, for attention, we coded chimpanzees’ looking at the two
D. Buttelmann et al. / Animal Behaviour 86 (2013) 755e761
apparatuses during the test phase, measured by the orientation of
their head. We did so (1) whenever they were in the squares and (2)
whenever they were visible on the videotapes. For the imitation
test, we coded the first action (lift or slide) that subjects used on the
apparatus to obtain a reward, and which actions they used to open
the doors in the 12 trials in which they were tested.
For interobserver reliability a second coder, who was naïve to
the aims of the study and blind to condition and action used in the
information phase, scored a randomly chosen 28% of the subjects
(¼5 subjects) for proximity and attention, and 30% (¼3 subjects) for
the actions used in the imitation test. Agreement was excellent
(Spearman correlations: proximity: left square: rS ¼ 1.00,
P < 0.001; right square: rS ¼ 0.975, P ¼ 0.005; attention: left
apparatus when in square: rS ¼ 0.975, P ¼ 0.005; right apparatus
when in square: rS ¼ 0.900, P ¼ 0.037; left apparatus whenever
visible: rS ¼ 0.900, P ¼ 0.037; right apparatus whenever visible:
rS ¼ 1.00, P < 0.001) and for imitation Cohen’s Kappa ¼ 1.0.
To assess the effect of condition on proximity and attention, we
calculated the difference between the time spent in front of or
looking at the appropriate action (the one used in the information
phase) and the time spent in front of the inappropriate action for
each individual and then compared the conditions. All reported P
values are two tailed.
759
Imitation test
Even after chimpanzees’ extensive exposure to successful box
opening with either lifting or sliding, and even after they had
shown recognition of this action subsequently, the chimpanzees
still failed to imitate the appropriate action, both during their first
trial (five performed the appropriate action; five performed the
inappropriate action; Fisher’s exact test, P ¼ 1.00) and overall
(means: 7.1 trials appropriate, 4.9 trials inappropriate; Wilcoxon
test: Tþ ¼ 29.5, N ¼ 9 (1 tie), P ¼ 0.461).
Self-experience Condition
Despite their previous experience, when chimpanzees in the
self-experience condition were presented with the choice between
the two humans attempting either to lift or to slide the door in the
recognition test, these chimpanzees did not choose to approach
and stay near to the individual who was using the method they
themselves had used and knew to work (Wilcoxon test: Tþ ¼ 27.0,
N ¼ 8, P ¼ 0.234; Fig. 3). The chimpanzees’ looking times in this
condition mirror the results of the proximity measure: irrespective
of how their looking behaviour was measured (see section on
Coding and Analyses) they showed no preference for either action
(Wilcoxon tests: both Ps 0.352).
Comparison Between Conditions
RESULTS
Modelled Condition
Mean time (s) spent in front of each action
Recognition test
Chimpanzees in the modelled condition spent more than twice
as much time in front of the action they had previously seen being
successful (Wilcoxon test: Tþ ¼ 48.0, N ¼ 10, P ¼ 0.033, r ¼ 0.66;
Fig. 3), thus demonstrating that they could recognize the particular
action being performed. Nine of the 10 individuals spent more time
in front of the appropriate action. The same results were found
when we analysed chimpanzees’ looking behaviour. Irrespective of
whether it was measured only when they were in the square(s) or
whenever they were visible on the screen, they looked at the
appropriate action longer (in square(s): 15 s; overall: 17 s) than at
the inappropriate action (6 s and 7 s, respectively; Wilcoxon tests:
þ
þ
¼ 41:0, N ¼ 10, P ¼ 0.028, r ¼ 0.69; Ttotal
¼ 48:5, N ¼ 10,
Tsquare
P ¼ 0.032, r ¼ 0.68).
40
*
30
Appropriate action
Inappropriate action
*
NS
20
Regarding the proximity measure, chimpanzees spent significantly less time in front of the appropriate action in the selfexperience condition than in the modelled condition in the
recognition test (ManneWhitney U test: U ¼ 17.5, Nmodel ¼ 10, Nselfexperience ¼ 8, P ¼ 0.045, r ¼ 0.20; Fig. 3). The same results were
found for the measure of subjects’ preference in looking time at the
two actions. Chimpanzees in the modelled condition (mean preference scores: in square(s): 9 s; overall: 10 s) showed a stronger
preference for the appropriate action than chimpanzees in the selfexperience condition (mean preference scores: 3 s and 3 s,
respectively), irrespective of when their looking behaviour was
measured (ManneWhitney U tests: Usquares ¼ 17.0, Nmodel ¼ 10,
Nself-experience ¼ 8, P ¼ 0.043, r ¼ 0.48; Utotal ¼ 14.5, Nmodel ¼ 10,
Nself-experience ¼ 8, P ¼ 0.021, r ¼ 0.54).
Additional Analyses
Additional analyses revealed that other factors cannot account for
our results. First, when we compared the time subjects spent in front
of the two different actions (lift versus slide), across condition and
action used in the information phase, subjects did not have a preference for one of the two actions (Wilcoxon test: Tþ ¼ 117.5, N ¼ 18,
P ¼ 0.171). Second, subjects’ imitation of the appropriate action was
not linked to the amount of time subjects had spent in front of or
looking at the appropriate action at test (Spearman correlations:
proximity: rS ¼ 0.276, N ¼ 10, P ¼ 0.440; attention: rS squares ¼ 0.485,
N ¼ 10, P ¼ 0.156; rS total ¼ 0.522, N ¼ 10, P ¼ 0.122).
DISCUSSION
10
0
25
12
11
Modelled
18
Self-experience
Condition
Figure 3. The mean number of seconds chimpanzees spent in front of each action in
the two conditions. Vertical lines indicate SEM. *P < 0.05.
Following a human’s demonstration of how to open the door of
a box successfully to obtain a food reward, chimpanzees showed
that they had paid attention to the particular bodily action the
human had used, and were able to remember and recognize it later,
when presented with two new humans manipulating the boxes in
different ways. Strikingly, however, when then given the opportunity to open the same box themselves, chimpanzees showed no
evidence of imitation (although they did show evidence of
emulation). These results help to pinpoint differences in the social-
760
D. Buttelmann et al. / Animal Behaviour 86 (2013) 755e761
learning abilities of chimpanzees and humans. Whereas human
children and adults readily translate their recognition of the successful actions of others into their own problem solving (Call &
Carpenter 2003; Wood et al. 2013), chimpanzees do not readily
perform this ‘from-you-to-me’ correspondence. The current results
demonstrate that this is not because they cannot or do not attend to
the actions of a demonstrator; they clearly attended to the actions
and, later, recognized the previously successful one. These results
further narrow down possible mechanisms that enable some
chimpanzees to copy some bodily actions after extensive training
(Hayes & Hayes 1952; Custance et al. 1995). Apparently it is not just
subjects’ attention that is trained (since our untrained subjects
readily paid attention to demonstrated actions) but other factors
that produce a change in the subjects during training. One possible
candidate might be subjects’ motivation to imitate (for a discussion
of social motivations possibly causing imitation in human children
see Over & Carpenter 2012; see also e.g. Myowa-Yamakoshi &
Matsuzawa 1999 and Call 2001 for other work exploring factors
that affect imitation performance).
Another important finding, from the self-experience condition,
was that even after the chimpanzees had used an instrumental
action successfully themselves (and thus had it in their motor
repertoire), they did not then show behavioural evidence on any of
our measures of recognizing when another individual was using
that same instrumental action. Chimpanzees’ failure to establish
this ‘from-me-to-you’ correspondence is surprising because a socalled mirror neuron system, which presumably also exists in
apes, should be enough to establish this correspondence (Rizzolatti
& Craighero 2004). It is also surprising given findings with human
infants that suggest that self-experience should support perception
and interpretation of others’ behaviour (Sommerville et al. 2005;
Meltzoff & Brooks 2008; Cannon et al. 2012).
One might wonder why chimpanzees in the modelled condition
did not systematically imitate in this study, when they did imitate,
or at least emulate, in the study by Horner et al. (2006). A very
similar apparatus was used in both studies, so differences in task
difficulty cannot explain the differences in imitative behaviour. Still,
there are several differences in the procedures of the two studies
that might explain this. For example, a human demonstrator was
used in the current study whereas a chimpanzee demonstrator was
used in the Horner et al. study. It might be reasonable to suppose
that chimpanzees would be more likely to pay attention to and
imitate conspecifics than humans. However, the chimpanzees
tested in the current study have been shown to pay attention to
humans in a wide variety of experiments (e.g. Call et al. 2004; Hare
& Tomasello 2004; Buttelmann et al. 2008, 2009, 2012), as well as
the current one, and have even socially learned from human
demonstrators in other, different types of tasks (e.g. Tennie et al.
2010). Furthermore, if one expected different results in our study
with conspecific models one would need to explain why, in other
studies, chimpanzees also fail to copy their groupmates (e.g.
Tomasello et al. 1997; Tennie et al. 2012).
More importantly, the study of Horner et al. (2006) was not
designed to test copying of specific actions, but only different ways
that a box operated (i.e. they did not look at exactly how subjects
used their hands, for example), and other studies looking for the
imitation of specific actions have revealed negative results (e.g.
Tennie et al. 2006). The current study was also about manipulating
an apparatus in different ways, but because of our experimental
design the apes really needed to pay attention to the specific action
being performed, not just the movement of the apparatus, in the
recognition phase. They succeeded, which shows that they can pay
attention to specific actions in some situations. However, they still
failed to perform those actions later, and even failed to reproduce
the box’s movements by emulation. One might argue that the
intervening recognition phase in our study somehow distracted the
subjects from even just emulating the box’s movements. However,
following the brief single trial of the recognition phase chimpanzees were presented with 12 further successful demonstrations
(with rewards) immediately prior to the imitation test. Thus, the
total number of successful actions far exceeded the number of
unsuccessful actions observed, so it seems unlikely that this brief
experience by itself would have affected their imitation. One might
also argue that chimpanzees’ imitation rate could have been
reduced by the fact that they had observed two actions preceding
the imitation task: the one performed by the original experimenter
and the other performed by one of the other two experimenters
during the recognition phase. However, we think this is also not
very plausible because the actions in the recognition phase were
only unsuccessful attempts to open the boxes and, therefore, were
not performed completely and were never associated with any
reward. Thus, the only action ever demonstrated to subjects
completely and associated with the retrieval of a reward was the
action demonstrated by the original experimenter. Whatever might
have caused our chimpanzees’ failure to imitate, the current study
shows that chimpanzees can pay attention to a demonstrator’s
actions, but still fail to imitate them, a finding that contributes
important information to the debate about how and why chimpanzees’ imitative skills differ from those of humans.
Imitation is a complex process with multiple components, and it
would seem that, among apes, humans are either especially skilful at
making correspondences between their own and others’ actions in
problem-solving situations, or especially motivated to do so. Our
results do not mean that chimpanzees are unable to copy the actions
of others under any circumstances. However, they suggest that an
enhanced ability or motivation to translate perceived actions into
motor output may have been a key component in Homo sapiens
evolving the ability for cultural transmission of both instrumental
and conventional actions with high fidelity across generations.
Acknowledgments
We thank Katja Grosse and Christian Nawroth for help conducting this study, Raik Pieszek for constructing the apparatuses,
Sylvio Tüpke for preparing Figs 1 and 2, and Lisa Sellge and Suska
Nolte for their help with coding the data.
References
ASAB/ABS 2012. Guidelines for the treatment of animals in behavioural research
and teaching. Animal Behaviour, 83, 301e309, http://dx.doi.org/10.1016/
j.anbehav.2011.10.031.
Boyd, R., Richerson, P. J. & Henrich, J. 2011. The cultural niche: why social learning
is essential for human adaptation. Proceedings of the National Academy of Sciences, U.S.A., 108, 10918e10925, http://dx.doi.org/10.1073/pnas.1100290108.
Buttelmann, D., Call, J. & Tomasello, M. 2008. Behavioral cues that great apes use
to forage for hidden food. Animal Cognition, 11, 117e128, http://dx.doi.org/
10.1007/s10071-007-0095-2.
Buttelmann, D., Call, J. & Tomasello, M. 2009. Do great apes use emotional expressions to infer desires? Developmental Science, 12, 688e698, http://
dx.doi.org/10.1111/j.1467-7687.2008.00802.x.
Buttelmann, D., Schütte, S., Carpenter, M., Call, J. & Tomasello, M. 2012. Great
apes infer others’ goals based on context. Animal Cognition, 15, 1037e1053,
http://dx.doi.org/10.1007/s10071-012-0528-4.
Caldwell, C. A., Schillinger, K., Evans, C. L. & Hopper, L. M. 2012. End state copying
by humans (Homo sapiens): implications for a comparative perspective on cumulative culture. Journal of Comparative Psychology, 126, 161e169, http://
dx.doi.org/10.1037/a0026828.
Call, J. 2001. Body imitation in an enculturated orangutan (Pongo pygmaeus).
Cybernetics and Systems, 32, 97e119.
Call, J. & Carpenter, M. 2003. On imitation in apes and children. Infancia y
Aprendizaje, 26, 325e349.
Call, J., Hare, B., Carpenter, M. & Tomasello, M. 2004. ‘Unwilling’ versus ‘unable’:
chimpanzees’ understanding of human intentional action. Developmental Science, 7, 488e498, http://dx.doi.org/10.1111/j.1467-7687.2004.00368.x.
D. Buttelmann et al. / Animal Behaviour 86 (2013) 755e761
Cannon, E. N., Woodward, A. L., Gredebäck, G., von Hofsten, C. & Turek, C. 2012.
Action production influences 12-month-old infants’ attention to others’ actions.
Developmental
Science,
15,
35e42,
http://dx.doi.org/10.1111/j.14677687.2011.01095.x.
Custance, D. M., Whiten, A. & Bard, K. A. 1995. Can young chimpanzees (Pan
troglodytes) imitate arbitrary actions? Hayes and Hayes (1952) revisited.
Behaviour, 132, 837e859.
Dean, L. G., Kendal, R. L., Schapiro, S. J., Thierry, B. & Laland, K. N. 2012. Identification of the social and cognitive processes underlying human cumulative
culture. Science, 335, 1114, http://dx.doi.org/10.1126/science.1213969.
Galef, B. G. & Laland, K. N. 2005. Social learning in animals: empirical studies and
theoretical models. Bioscience, 55, 489e499, http://dx.doi.org/10.1641/00063568(2005)055[0489:SLIAES]2.0.CO;2.
Hare, B. A. & Tomasello, M. 2004. Chimpanzees are more skillful in competitive
than in cooperative cognitive tasks. Animal Behaviour, 68, 571e581, http://
dx.doi.org/10.1016/j.anbehav.2003.11.011.
Haun, D. B. M. & Call, J. 2008. Imitation recognition in great apes. Current Biology,
18, R288eR290, http://dx.doi.org/10.1016/j.cub.2008.02.031.
Hayes, K. J. & Hayes, C. 1952. Imitation in a home-raised chimpanzee. Journal of
Comparative and Physiological Psychology, 45, 450e459.
Hopper, L. M., Spiteri, A., Lambeth, S. P., Schapiro, S. J., Horner, V. & Whiten, A.
2007. Experimental studies of traditions and underlying transmission processes
in chimpanzees. Animal Behaviour, 73, 1021e1032, http://dx.doi.org/10.1016/
j.anbehav.2006.07.016.
Hopper, L. M., Lambeth, S. P., Schapiro, S. J. & Whiten, A. 2008. Observational
learning in chimpanzees and children studied through ‘ghost’ conditions. Proceedings of the Royal Society B, 275, 835e840, http://dx.doi.org/10.1098/
rspb.2007.1542.
Horner, V. & Whiten, A. 2005. Causal knowledge and imitation/emulation
switching in chimpanzees (Pan troglodytes) and children (Homo sapiens). Animal
Cognition, 8, 164e181, http://dx.doi.org/10.1007/s10071-004-0239-6.
Horner, V., Whiten, A., Flynn, E. & de Waal, F. B. M. 2006. Faithful replication of
foraging techniques along cultural transmission chains by chimpanzees and
children. Proceedings of the National Academy of Sciences, U.S.A., 103, 13878e
13883, http://dx.doi.org/10.1073/pnas.0606015103.
Itakura, S., Agnetta, B., Hare, B. & Tomasello, M. 1999. Chimpanzee use of human
and conspecific social cues to locate hidden food. Developmental Science, 2,
448e456, http://dx.doi.org/10.1111/1467-7687.00089.
Kendal, R. L., Coolen, I., van Bergen, Y. & Laland, K. N. 2005. Tradeoffs in the adaptive
use of social and asocial learning. Advances in the Study of Behavior, 35, 333e379.
LaBarbera, J. D., Izard, C. E., Vietze, P. & Parisi, S. A. 1976. Four- and six-month-old
infants’ visual responses to joy, anger, and neutral expressions. Child Development, 47, 535e538.
Laland, K. N. 2004. Social learning strategies. Learning and Behavior, 32, 4e14,
http://dx.doi.org/10.3758/BF03196002.
Meltzoff, A. N. & Brooks, R. 2008. Self-experience as a mechanism for learning
about others: a training study in social cognition. Developmental Psychology, 44,
1257e1265, http://dx.doi.org/10.1037/a0012888.
Munakata, Y., Santos, L. R., Spelke, E. S., Hauser, M. D. & O’Reilly, R. C. 2001. Visual
representation in the wild: how rhesus monkeys parse objects. Journal of
Cognitive Neuroscience, 13, 44e58, http://dx.doi.org/10.1162/089892901564162.
Myowa-Yamakoshi, M. & Matsuzawa, T. 1999. Factors influencing imitation of
manipulatory actions in chimpanzees. Journal of Comparative Psychology, 113,
128e136, http://dx.doi.org/10.1037/0735-7036.113.2.128.
761
Nagell, K., Olguin, K. & Tomasello, M. 1993. Processes of social learning in the tool
use of chimpanzees and human children. Journal of Comparative Psychology, 107,
174e186.
Nielsen, M., Collier-Baker, E., Davis, J. M. & Suddendorf, T. 2005. Imitation
recognition in a captive chimpanzee (Pan troglodytes). Animal Cognition, 8,
31e36, http://dx.doi.org/10.1007/s10071-004-0232-0.
Over, H. & Carpenter, M. 2012. Putting the social into social learning: explaining
both selectivity and fidelity in children’s copying behavior. Journal of Comparative Psychology, 126, 182e192, http://dx.doi.org/10.1037/a0024555.
Paukner, A., Anderson, J. R., Borelli, E., Visalberghi, E. & Ferrari, P. F. 2005.
Macaques (Macaca nemestrina) recognize when they are being imitated. Biological Letters, 1, 219e222.
Paukner, A., Suomi, S. J., Visalberghi, E. & Ferrari, P. F. 2009. Capuchin monkeys
display affiliation toward humans who imitate them. Science, 325, 880e883,
http://dx.doi.org/10.1126/science.1176269.
Pfungst, O. 1911. Clever Hans (The Horse of Mr. von Osten): a Contribution to
Experimental Animal and Human Psychology. Translated by C. L. Rahn. New York:
Henry Holt.
Rizzolatti, G. & Craighero, L. 2004. The mirror-neuron system. Annual Review of
Neuroscience, 27, 169e192, http://dx.doi.org/10.1146/annurev.neuro.27.070203.
144230.
Sommerville, J. A., Woodward, A. L. & Needham, A. 2005. Action experience alters
3-month-old infants’ perception of others’ actions. Cognition, 96, B1eB11,
http://dx.doi.org/10.1016/j.cognition.2004.07.004.
Tennie, C., Call, J. & Tomasello, M. 2006. Push or pull: emulation versus imitation
in great apes and human children. Ethology, 112, 1159e1169, http://dx.doi.org/
10.1111/j.1439-0310.2006.01269.x.
Tennie, C., Call, J. & Tomasello, M. 2009. Ratcheting up the ratchet: on the evolution of cumulative culture. Philosophical Transactions of the Royal Society B,
364, 2405e2415, http://dx.doi.org/10.1098/rstb.2009.0052.
Tennie, C., Call, J. & Tomasello, M. 2010. Evidence for emulation in chimpanzees in
social settings using the floating peanut task. PLoS One, 5, e10544, http://
dx.doi.org/10.1371/journal.pone.0010544.
Tennie, C., Call, J. & Tomasello, M. 2012. Untrained chimpanzees (Pan troglodytes
schweinfurthii) fail to imitate novel actions. PLoS One, 7, e41548, http://
dx.doi.org/10.1371/journal.pone.0041548.
Tomasello, M. 1996. Do apes ape? In: Social Learning in Animals: the Roots of Culture
(Ed. by J. Galef & C. Heyes), pp. 319e346. New York: Academic Press.
Tomasello, M. 1999. The Cultural Origins of Human Cognition. Cambridge, Massachusetts: Harvard University Press.
Tomasello, M., Call, J. & Gluckman, A. 1997. Comprehension of novel communicative signs by apes and human children. Child Development, 68, 1067e1080,
http://dx.doi.org/10.2307/1132292.
Whiten, A., Horner, V. & Marshall-Pescini, S. 2003. Cultural panthropology.
Evolutionary Anthropology, 12, 92e105, http://dx.doi.org/10.1002/evan.10107.
Whiten, A., Horner, V. & de Waal, F. B. M. 2005. Conformity to cultural norms of
tool use in chimpanzees. Nature, 437, 737e740, http://dx.doi.org/10.1038/
nature04047.
Williamson, R. A., Donohue, M. R. & Tully, E. C. 2013. Learning how to help others:
two-year-olds’ social learning of a prosocial act. Journal of Experimental Child
Psychology, 114, 543e550, http://dx.doi.org/10.1016/j.jecp.2012.11.004.
Wood, L. A., Kendal, R. & Flynn, E. 2013. Copy you or copy me? The effect of prior
personally-acquired, and alternative method, information on imitation. Cognition, 127, 203e213, http://dx.doi.org/10.1016/j.cognition.2013.01.002.