Expert answer:Question Discussion

Expert answer:1. Which ability do the ravens appear to demonstrate?2. How does the ravens’ ability compare to similar ability in chimpanzees? (Look back at your notes for Call & Tomasello, 2008.)3.Does the ravens’ ability appear to be human-like, or different?
bugnyar_et_al.__2015_.pdf

Unformatted Attachment Preview

ARTICLE
Received 29 Jul 2015 | Accepted 18 Dec 2015 | Published 02 Feb 2016
DOI: 10.1038/ncomms10506
OPEN
Ravens attribute visual access to unseen
competitors
Thomas Bugnyar1,2, Stephan A. Reber1,2 & Cameron Buckner3
Recent studies purported to demonstrate that chimpanzees, monkeys and corvids possess a
basic Theory of Mind, the ability to attribute mental states like seeing to others. However,
these studies remain controversial because they share a common confound: the conspecific’s
line of gaze, which could serve as an associative cue. Here, we show that ravens Corvus corax
take into account the visual access of others, even when they cannot see a conspecific.
Specifically, we find that ravens guard their caches against discovery in response to the
sounds of conspecifics when a peephole is open but not when it is closed. Our results suggest
that ravens can generalize from their own perceptual experience to infer the possibility of
being seen. These findings confirm and unite previous work, providing strong evidence that
ravens are more than mere behaviour-readers.
1 Department of Cognitive Biology, University of Vienna, Althanstrasse 14, 1090 Wien, Austria. 2 Haidlhof Research Station, University of Vienna and
University of Veterinary Medicine, 2540 Bad Vöslau, Vienna. 3 Department of Philosophy, University of Houston, 4300 Calhoun Road, Houston, Texas 77004,
USA. Correspondence and requests for materials should be addressed to T.B. (email: thomas.bugnyar@univie.ac.at).
NATURE COMMUNICATIONS | 7:10506 | DOI: 10.1038/ncomms10506 | www.nature.com/naturecommunications
1
ARTICLE
K
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10506
nowing what others see would provide animals with an
advantage when competing for food, for it would allow
them to predict which items might become the subject of
disputes. In line with this ecologically based assumption,
behavioural experiments manipulating the visibility of food in
face-to-face competition tasks have revealed that non-human
primates1–3 and corvids4–6 react to what conspecifics can and
cannot see. Many researchers have thus concluded that these
animals possess an understanding of perception-goal
psychology—a basic ‘Theory of Mind’. However, because in
most of these experiments successful performance can be
achieved by tracking correlations between head cues and a
competitor’s behaviour, skeptics have concluded that all
these experiments suffer from a ‘logical problem’ that
renders them unable to empirically distinguish representations
of directly observable cues from a genuine representation of
‘seeing’7–10.
In an attempt to overcome this problem, researchers have tried
to develop designs that control for others’ gaze. In a landmark
study, Emery and Clayton4 showed that scrub-jays recache food
after being watched at caching. The design controlled for
occurrent gaze cues in the test by blocking visual access to the
competitor during recaching. However, it did not control for a
memory of past gaze cues, as the competitor was present during
the initial caching episode—and so even this elegant design can
be explained by appeal to representations of previously observed
gaze cues alone7,8. Recently, Schmelz et al.11,12 showed that
chimpanzees predict others’ preferences in a back-and-forth
foraging game even when they never saw their competitors’ gaze
at the stimuli used in the test, and Ostojić et al.13,14 found that
Eurasian jays can predict the food preferences of their mates only
if they see what they have been prefed, even if denied prior visual
access to the partners’ response to prefeeding in this context.
Although these lines of research both control for behavioural cues
in sophisticated ways, they focus on attributions of preferences
rather than sight. Thus, it still remains an open question whether
any nonhuman animal can attribute the concept ‘seeing’ without
relying on behavioural cues.
Like scrub-jays15, ravens also cache food items spontaneously
and they are highly sensitive to the presence of conspecifics that
may pilfer caches16. In particular, ravens decrease the likelihood
of revealing cache locations to competitors by (i) reducing the
time to finish caches, (ii) using obstacles as visual barriers for
caching outside the view of competitors, (iii) delaying caching
until competitors have left and (iv) staying away from already
made caches as long as competitors are around17,18. Across all
these studies, however, another raven was visible during the test,
again raising the skeptical refrain that ravens might represent
only gaze cues10. Suggestively, however, Stulp et al.19 and Shaw
and Clayton20,21 have recently shown that jays can infer the
presence of unseen conspecifics on the basis of auditory cues.
These findings point the way towards the experimental design
reported here, which aims to finally overcome the line-of-gaze
interpretation by controlling for both (i) gaze cues in the test, by
relying only on sounds to indicate the presence of a possible
competitor and (ii) memories of gaze cues previously observed in
the training, by denying the subjects prior access to any
competitor’s gaze in contexts resembling the test. Specifically,
we ask if ravens can transfer knowledge from their own
experience in a novel context—using peepholes to look into an
adjacent room—to a caching situation in which they can hear but
not see a conspecific in that room.
The experiment was conducted in two rooms separated by a
wooden wall. The wall had two functional windows that allowed
visual access from one room into the other. Both windows could
be closed with covers. Furthermore, each window had a peephole
2
drilled into its cover, which could be independently opened or
closed (Supplementary Fig. 1). At the onset of the study, all ravens
were familiar with experimenters opening or closing the windows
between the rooms but were naı̈ve about the peepholes. In an
initial, baseline step, we determined the caching behaviour of
individual ravens when they received food in one room and the
windows to the adjacent room containing conspecifics were open
(observed condition) or closed (non-observed condition;
Supplementary Tables 2 and 3). Crucially, in both of these
conditions, the peepholes were closed. In the next, familiarization
step, each raven was individually trained to use one of the
peepholes to observe and recover human-made caches in the
adjacent room. In the final stage, each raven was confronted with
the two baseline conditions (windows opened or closed) and an
interspersed test condition, in which the windows between the
two rooms were closed, but one of the two peepholes was open
(Fig. 1). Rather than another raven being present in the test,
however, the observation room was empty, and a hidden
loudspeaker was used to play a series of sounds that were
recorded from a competitor raven in the observation chamber
during the non-observed condition.
The experimental hypothesis that ravens learn in the
familiarization trials that peepholes afford seeing predicts that
ravens will behave in the test as though they are potentially being
observed through the peephole. Specifically, from the findings in
the observed and non-observed conditions of the initial baseline
step, we would expect ravens to (i) finish caches more quickly in
the peephole condition than in the non-observed condition and
(ii) return to improve their caches less often in the peephole
condition than in the non-observed condition, and that their
caching behaviour in the peephole condition would not differ
a
Obs
st
b
Non
st
c
Peep
st
Figure 1 | Sketch of experimental setup. (a) Observed (Obs) condition:
The cover of the window is open (white bar) and the focal subject
(storer, st) caches food in the visual presence of a conspecific (observer).
(b) Non-observed (Non) condition: The cover of the window is closed
(grey bar) and the focal subject caches food in visual isolation of a
conspecific (non-observer). Both observers and non-observers make
sounds in the experimental chamber, which are audible to the storer.
(c) Peephole (Peep) condition: The cover of the window is closed (grey bar)
but one of the two peepholes (small white square) is open; the focal
subject caches food in the absence of any behavioural cues, whereas the
presence of conspecifics is simulated via playback of sounds recorded
from non-observed trials (symbolized by loudspeaker).
NATURE COMMUNICATIONS | 7:10506 | DOI: 10.1038/ncomms10506 | www.nature.com/naturecommunications
ARTICLE
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10506
significantly from their caching behaviour in the observed
condition.
We show that ravens treat the test condition like the observed
condition, indicating that they can generalize from their own
experience using the peephole as a pilferer and predict that
audible competitors could potentially see their caches.
Consequently, we argue that they represent ‘seeing’ in a way that
cannot be reduced to the tracking of gaze cues.
Results
Caching with peephole as though observed. Figure 2
summarizes the main findings. As expected, ravens differed
between conditions in the time to finish a cache (w2 ¼ 14.889,
d.f. ¼ 2, Po0.001) and the number of revisits to improve a cache
(w2 ¼ 12.929, d.f. ¼ 2, Po0.001). These data clearly support both
experimental hypotheses. The subjects finished their caches more
quickly and they returned to improve their caches less often in the
peephole condition than in the non-observed condition
(Wilcoxon signed-ranks test, Bonferroni corrected: time to finish:
n ¼ 9, Z ¼ 2.666, P ¼ 0.012; revisit and improve: n ¼ 9,
Z ¼ 2.539, P ¼ 0.024). No significant difference could be found
between the peephole and the observed condition regarding these
hypotheses (time to finish: n ¼ 9, Z ¼ 2.192, P ¼ 0.08; revisit
and improve: n ¼ 9, Z ¼ 1.414, P ¼ 0.94). Note that ravens did
go back to visually inspect their caches in several conditions
a
*
Average duration caching (s)
18
*
NS
16
14
12
10
Using peephole’s limited viewing angles to hide caches. The
experimental setup also allowed us to investigate whether ravens
understand more detailed information about the point of view of
possible observers. As there were few physical structures in the
caching room that could serve as visual barriers and observers
could freely move between the two windows, it was difficult for
the ravens to place caches outside of view of conspecifics in the
observed condition. In the peephole trials, however, our design of
opening one peephole at a time enabled us to use parts of the
front wall as a visual barrier. Hence, this design allowed us to test
whether ravens could predict which side of the room could not be
seen by a potential competitor from a particular peephole. If they
could, they should avoid caching in the area in front of the open
peephole only. Although particular individuals did place their
caches to the left when the right hole was open and to the right
when the left hole was open (see Supplementary Fig. 2 for details),
the ravens as a group did not show a significant preference for
caching away from the open peephole (1.78 caches made inside
view versus 2.78 caches made outside view, Wilcoxon signed
ranks test: n ¼ 9, Z ¼ 1.185, P ¼ 0.297). These results indicate that
ravens need visual feedback for effectively using barriers and
point towards possible limitations of their attribution skills.
8
6
Non
Obs
Peep
b
*
Average number improvements
(Friedman test: w2 ¼ 1.032, d.f. ¼ 2, P ¼ 0.625), indicating that
they were not generally inhibited in returning to their caches in
the observed and peephole conditions. This asymmetry makes
adaptive sense, as prior studies have shown that observer ravens
do not use gaze direction to find hidden food (despite being
skilled at following others’ gaze into distance and around barriers)
but are very prone to show enhancement when they see someone
touching an item22. Because all available behaviour-reading cues
have been controlled for in the test condition—there is no actual
competitor whose gaze could be read, and the situation is novel
from the subject’s perspective—these data provide clear evidence
that raven social cognition cannot be reduced to behaviourreading.
*
1.5
NS
1.0
0.5
0.0
Non
Obs
Peep
Figure 2 | Effects of condition on caching behaviour. (a) Mean time to
finish a cache and (b) mean number of revisits with improvements, in the
non-observed condition (Non, total of 4 trials per 10 ravens), observed
condition (Obs, total of 4 trials per 10 ravens) and peephole condition
(Peep, total of 2 trials per 9 ravens). Box plots represent 25th and 75th
percentiles, centre line indicates the median, whiskers represent non-outlier
range and dots are outliers (Friedman test, post hoc Wilcoxon signed ranks
test; *Po0.05; NS ¼ non-significant).
Discussion
Do the current data provide evidence that ravens have a Theory
of Mind? A difficulty is finding an empirical criterion for
assessing the presence of Theory of Mind that applies to the
current debate over attributions of seeing in nonlinguistic
animals. The most popular criterion for assessing Theory of
Mind in general has been the false belief task, suggested in
response to Premack and Woodruff’s original article on Theory of
Mind23. Following Wimmer and Perner’s application of this test
to children24, it has become the standard benchmark for Theory
of Mind development in humans25. However, this criterion is not
well-suited to arbitrate the current dispute, because the clearest
versions of the task rely on language and because it assesses the
attribution of more advanced epistemic states (like belief) rather
than perceptual states (like seeing). Indeed, most of the
researchers cited in the introduction concede that there is little
evidence that nonhuman animals possess the full-blown
metarepresentational capacities required to attribute false
beliefs, but still think there is an interesting empirical debate
to be had regarding whether any other social animals
have the perceptual precursors to these abilities (such as the
‘secondary representations’ posited in Perner’s developmental
model—see ref. 26).
The closest one finds to an ecumenical proposal that could
arbitrate this dispute is Whiten’s27,28‘intervening variable’
solution (see also ref. 7 for evidence that some skeptics accept
this benchmark). The current data provide evidence that ravens
satisfy this criterion. The core idea of Whiten’s proposal is that an
NATURE COMMUNICATIONS | 7:10506 | DOI: 10.1038/ncomms10506 | www.nature.com/naturecommunications
3
ARTICLE
NATURE COMMUNICATIONS | DOI: 10.1038/ncomms10506
animal represents an underlying mental state when it groups a
variety of causes and effects together under a common internal
code. As evidence for such a code, the ability to behave
appropriately in novel conditions by flexibly integrating
information from perceptually dissimilar situations is crucial.
Our peephole test condition is novel in this sense, and the data
concerning cache duration and number of improvements are
consistent with an intervening variable interpretation.
Because conspecifics were never present in the peephole
context in either the pretraining or the test conditions, these
data demonstrate a flexible ability to individually modulate
distinct behaviours to specific and novel circumstances, which
cannot be reduced to a tracking of gaze cues. Recent critical
discussions of the intervening variable solution have emphasized
just these forms of flexibility and integration, suggesting that the
explanatory power of a Theory of Mind hypothesis may lie in
the capacity of the animal to learn new observable cues
(such as open peepholes) to indicate the presence of a
competitor’s mental state9,10. The observed, non-observed and
peephole conditions are clearly perceptually dissimilar, and
cache speed and inhibition of revisits reflect skills that are
clearly distinct.
Moreover, no combination of individual stimulus-response
links can explain the full pattern of observed data. The presence
of peepholes alone cannot explain the results, because the subjects
lack a specific associative history for caching in the presence of
peepholes; as evidence that they did not simply regard them as a
familiar form of visual occlusion, note that they were initially
fearful of the peepholes in familiarization trials, and one raven
never achieved criterion performance. Neither can the presence of
the playback sounds alone explain the results, because if the
ravens could detect the difference between live and playback
stimuli, the latter should be perceived as strange and elicit
cautious behaviour in playback trials in general. However, the
ravens did not show any inhibition in caching or returning to
their caches in the test condition; they only inhibited behaviours
indicative of ‘being observed’, like the touching of already made
caches. This selectivity in response cannot be explained by
perceiving playback as a novel stimulus that elicits caution.
Finally, we should note that the current results cannot be
explained merely on the basis of stress or anxiety as suggested by
a computational model offered recently by Van der Vaart et al.29.
The stress hypothesis predicts increased caching when observed
and increased revisits, but the opposite was found in our study
(see also refs 30,31).
Skeptics might further worry that the effects observed in this
study were artefacts of human enculturation, given that the
ravens were human-reared and had experience interacting with
humans in behavioural experiments (as has sometimes been
suggested to explain the social cognition of dogs and apes—see
ref. 32). In short, the ravens might only have learned to use the
peepholes because they have acquired special learning abilities
because of the control that humans wield over their food supply.
Even if this were the case, it would not constitute a good
argument against the interpretation of the data offered
here—unless we were prepared to accept that Theory of Mind
is an artefact in humans as well. From birth, children are also
extraordinarily dependent upon their caregivers for food, and are
also entrained, implicitly and explicitly, to attune to the social
behaviours of their conspecifics. It is crucial in cross-species
comparisons that we apply the same yardstick to humans and
animals33. Thus, although it would still be of considerable
evolutionary interest to determine whether parent-raised ravens
could pass the peephole test without human enculturation, this
worry should not lead us to reject the current interpretation of the
results.
4
It may also be promising to consider the present findings in
terms of the ‘minimal’ (as opposed to ‘full-blown’) Theory of
Mind recently articulated by ref. 34. Agents with minimal Theory
of Mind adaptively respond to mental states by representing
‘encounterings’, defined as relations between agents and objects in
their visual field. Such animals can learn that ‘having encountered
food (is a) a condition for performing goal-directed actions
targeting that food’, and can thus, to prevent theft, be motivated
to prevent others from encountering their caches. To
accommodate our data, this principle must be elaborated to
include the computation of possible encounterings by agents even
when no competitor is visible and when generalized from their
own perceptual experience. This may appear to approach some of
the most sophisticated criteria proposed for perceptual Theory of
Mind, such as Heyes’ ‘projection’ test10. However, at least as a
group the ravens’ ability fell short of ‘full-blown’ human Theory
of Mind, as it did not enable them to anticipate the limited
viewing angle from one peephole or another—although this is a
sophis …
Purchase answer to see full
attachment

How it works

1. Paste your instructions in the instructions box. You can also attach an instructions file
2. Select the writer category, deadline, education level and review the instructions 
3. Make a payment for the order to be assignment to a writer
4.  Download the paper after the writer uploads it 

Will the writer plagiarize my essay?

You will get a plagiarism-free paper and you can get an originality report upon request.

Is this service safe?

All the personal information is confidential and we have 100% safe payment methods. We also guarantee good grades