that future planning in corvids is a cogni-
tive ability that can only be used during
food caching or some other specific evo-
lutionarily selected adaptation. The study
provides evidence that future planning can
flexibly serve different behaviors not only in
humans but also in nonhuman animals.
The authors presented five ravens with
a choice of objects. Only one of these ob-
jects was a functional tool, which could be
used to retrieve food from a puzzle box.
The ravens chose correctly not only when
they were offered the box but also when
they had to store the tool and
plan for the next day. In another experiment, the ravens
were trained to exchange tokens for food. When the ravens
knew that trading would only
happen on the next day, they
chose and stored these tokens
as soon as they were offered
to them. By manipulating tool
choice, time, and trading opportunities, the authors controlled the value of the items at
choice in relation to current as
well as future interactions.
The results from the two experiments show that ravens
take temporal distance between
item choice and reward into account, exercise self-control, and
make decisions for predicted
futures rather than arbitrary
ones. Thus, the birds opt for a
more distant but higher gratification rather than an immediate but lower gratification and
do so flexibly across behaviors.
What are the selective pres-
sures that led to complex cog-
nition and enabled future
planning to evolve? The answer
to this question may be best
understood by investigating
convergent evolution among
distantly related species such as corvids
and primates (4). In the 1990s, scientists
discovered that mammals and birds share
homologous forebrain structures, the neu-
robiological foundation for complex cog-
nition (7). On the basis of this evidence,
corvids and parrots were tested for cogni-
tive abilities to explore the evolutionary
development of cognition. There are two
main hypotheses to explain the evolution of
intelligence. According to the physical intel-
ligence hypothesis, complex cognitive ca-
pabilities evolved according to the physical
demands of the environment, such as the
need to memorize location, time, and avail-
ability of food or how it can be extracted
with tools. The social intelligence hypoth-
esis focuses on the social environment and
the requirement to anticipate and manipu-
late the behavior of individuals from the
same and other species. These mechanisms
need not be mutually exclusive and might
interact with phylogenetic heritage and de-
velopmental constraints (8).
The evolution of future planning in corvids may provide a good example of how
these selective pressures are entwined.
Corvids commonly cache perishable foods,
which they only consume while still fresh.
They must therefore understand the physical
features of what they cached and when and
where they did so. On the other hand, these
birds are social; they can follow the gaze of
others, remember where others have cached,
and use their observational memory to pilfer
those caches later. They also tactically deceive conspecifics who try to pilfer their food
caches. Ravens cache food from ephemeral
but temporarily and locally highly abundant
carcasses. They use several tactics to protect
their caches from being stolen by others as
well as pilfer those that others have cached
and anticipate their competitors’ behavior
(9). Corvids thus evolved complex cognitive
abilities to deal with the social and physical
environment but use their intelligence flexibly across situations.
Future planning in corvid food caching
seems to be controlled by two different motivational systems: a feeding system and
a caching system. One of us (N.S.C.) has
studied these systems in an experiment in
which individual corvids were fed with peanuts, powdered peanuts, and/or peanut-sized stones (10). Subsequently, the birds
were presented with whole peanuts or
stones, and their caching behavior was observed. Only a combination of motivational
changes due to feeding as well as caching in the previous stage explained later
caching of stones and peanuts.
Food storing thus seems to be
controlled by the relatively
independent caching system
interacting with the feeding
system, jointly controlling the
motivation to cache and plan
for the future.
In humans, multiple motivations are at work when planning for the future. Kwan et
al. have argued that there are
two simultaneous but different
cognitive processes involved in
future planning in humans; one
is directly related to the value of
a future reward and the other
to imagining the experience of
a future reward (11). “Never go
shopping when you are hungry”
is a common advice, but overbuying might be reduced when
specifically imagining oneself
in the future. By dissociating
current and future value, adequate performance in future
planning increases (11). Ravens
and other food-caching corvids
might similarly use two potential routes for future planning—
namely, orienting to the value
of future reward and/or imagining themselves retrieving food
caches in the future. For ravens
just as for humans, memories are thus more
for the future than for the past. j
1. T. Suddendorf, D. R. Addis, M. C. Corballis, Philos. Trans.R.
Soc. B 364, 1317 (2009).
2. J.Redshaw, T.Suddendorf, Front. Psychol. 4,404(2013).
3. C. R. Raby, D. M. Alexis, A. Dickinson, N. S. Clayton, Nature
445, 919 (2007).
4. N. J. Emery, N. S. Clayton, Science306, 1903 (2004).
5. C. Kabadayi, M. Osvath,Science 357, 202 (2017).
6. N. S. Clayton, T. J. Bussey, A. Dickinson, Nat. Rev. Neurosci.
4, 685 (2003).
7. E.Jarvis et al., Nature 6, 151(2005).
8. K.E.Holekamp, Tr. Cogn. Sci. 11,65(2007).
9. T.Bugnyar, S.A.Reber,C.Buckner, Nat. Commun. 7,10506
10. N. S. Clayton, A. Dickinson, Anim.Behav. 57, 435 (1999).
11. D. Kwan et al., Hippocampus 22, 1215 (2012).
Tool-using raven. In a separate experiment, a raven involved in the Kabadayi and
Osvath study uses a stick to poke for food in a tube (first photo). She succeeds in
pushing the stick into the hole (second photo), but handling the stick is tricky for
her. She then invents a new way of solving the problem: She fills the tube with bark
pieces (third photo) and thereafter pecks at the bark until the food falls out.