that this is the very worst time of day for such a traumatic event to take place.
Not only would elephant herds within a radius of at least 10 km (6 mile) hear
the sounds of distress and panic of their fellow species, but they would hear the
infrasound from helicopter blades and gunfire over distances of perhaps greater
than 100 km (60 mile). They would thus be all but witness to this traumatic
scene. Payne and Martin (personal communication, 1995) report that in a cull-
ing operation in the Sengwa National Park in Zimbabwe, elephants hearing the
sounds of the operation fled 145 km (90 mile) to the borders of the park.
That elephants can detect the seismic field of the earth and use this signal to
navigate remains a possibility that awaits further research.
The huge ears of the African savanna elephant hearing system also serve the
function of dissipating heat in environments where daytime temperatures can
exceed 40 °C (104 °F). Heat loss through the blood vessels near the surface of
the ears is increased by flapping of the ears. For animals weighing as much as
7 metric tons, however, this may not be enough. Recent work has shown that
other areas of the elephant’s body contain dense networks of blood vessels to
which the flow of blood may be selectively controlled and used as locations
for heat dissipation. Weissenbock and her team at the Vienna Zoo have shown
surface temperature differences on the elephant’s body to differ by as much
as 20 °C (68 °F). They suggest that elephants may be able to consciously open
and close these thermal windows (Weissenböck et al., 2010, 2012). Other work
being conducted at Princeton University by Myhrvold suggests that the sparse
covering of bristly hairs found on an elephant might act as “tiny heat fins.” A
computer model developed by Myhrvold to test this idea suggests that hairs can
boost the heat loss from an elephant’s body by 20% (Myhrvold et al., 2012).
Elephants may also take advantage of the extremely low temperatures ex-
perienced on the open dry tropical savannas at night. As decribed earlier in
the Introduction, low humidity and cloudless skies in savanna elephant habitat
72 Elephant Sense and Sensibility
result in huge and rapid radiative losses of heat from the earth’s surface to space.
Surface temperatures may drop from daytime highs of above 40 °C (104 °F) to
less than 5 °C (41 °F) beginning before sunset. Elephants may allow their body
temperature to drop abnormally low under these cold nighttime conditions in
order to cope with the extremely high daytime temperatures.
The demands made on the brain to process the amount and complexity of
auditory and other signals is considerable. Not only are a multitude of signals
taken in by the neural network, but these signals must be processed simultane-
ously in a variety of ways. Memory sources need to be combined to identify
acoustic sources with complex discriminatory functions that must determine the
identity and location of these sources. These operations are further extended to
incorporate rudimentary language.
Pijanowski and colleagues (2011), as well as previous researchers (Krauss,
1987; Southworth, 1969), have suggested that sound is fundamental to the eco-
logical landscape and as such must be incorporated as an essential part of the
environment on that scale. The value of this vision is that sound in the animal
world is much more than communication between and among species. Rather,
sound from nonbiological sources may be actively used in decision making
(movements), area occupied (home range), navigation (migration, resources),
and avoidance of danger (fire, flood) (Pijanowski et al., 2011).
Chapter 10
Language
We have suggested that elephants communicate with each other, perhaps using
a far more complex system than we do, incorporating sound, smell, sight, taste,
and touch. The question is whether any of this exchange translates into what
could be called language with a vocabulary and structure. Webster’s Collegiate
Dictionary, for example, defines language as “any means, vocal or other, of
expressing or communicating feeling or thought.” There is little doubt that
elephant communication meets such a definition.
In elephants, at least 10 different vocalization classes have been identified
(Clemins et al., 2005). Using speech processing techniques and seven captive
adult African elephants, Clemins and colleagues (2005) were able to identify
each of these vocalization types with an accuracy of between 80% and 94%
depending on the quality of the dataset used. The elephant emitting the sounds
could be recognized nearly 9 out of every 10 times.
Poole (2011) recognizes 13 call types divided into three groups based on the
source of the sound (Table 10.1). Seven rumble calls are classified as laryngeal calls, three as trunk calls, and three as initiated or idiosyncratic calls. The calls
may have a range of meanings, which Poole calls “gradiness,” as made by the
caller but may have a discrete meaning as interpreted by the receiver. Poole rec-
ognizes 35 contextual calls, many of which have a very discrete meaning such
as the “let’s go” rumble. She divides these 35 calls into eight contextual classes.
Poole (2011) indicates that there might be a range of meanings to the call
emitted but a discrete interpretation is made by the receiver. Defensive calls may
result in immediate reaction by all members of the herd gathering in a tightly
knit group or simply responding to where two herds are in contact with each
other. Payne (1998), for example, reports that GPS-collared groups feeding in
the same area and tracked over a number of weeks never crossed each other’s
paths. Such behavior avoids expenditure of energy in going into an area already
depleted in resources.
Communication between elephants over considerable distances indicates the
ability to locate the caller with considerable accuracy. Playback experiments of
female estrous calls (Langbauer et al., 1991) demonstrate that males can locate
the source of the recorded call 2 km (1.2 mile) away. Palmer (2004) points out
that while elephants are capable of using interaural time delays to determine the
location of calls, this process requires considerable neural processing. A network
Elephant Sense and Sensibility. http://dx.doi.org/10.1016/B978-0-12-802217-7.00010-7
Copyright © 2015 Elsevier Inc. All rights reserved.
73
74 Elephant Sense and Sensibility
TABLE 10.1 Sources of Elephant Sounds
Laryngeal Calls
1
Rumble
2
Rev
3
Roar/roar rumble
Tonal roar
Noisy roar
Mixed roar
Pulsated roar
4
Cry/cry rumble
5
Bark/bark rumble
6
Grunt
7
Husky cry
Trunk Calls
8
Trumpet
Pulsated trumpet
9
Nasal trumpet
10
Snort
Imitated and Idiosyncratic
11
; Truck-like
12
Croak
13
Squelch
Context Calls
Multiple subcall types related to:
Group defense
Food
Sexual
Mother–offspring
Conflict
Social integration
Logistical calls
Play
Modified after Poole, 2011.
Language Chapter | 10 75
of neurons must act in the elephant’s brain to detect the moment when the signal
is received simultaneously in each ear, identifying the direction probably within
1° of azimuth of the location of the source. Depending on the frequency of the
sound being processed by the brain, it is possible that further processing involves
phase changes in the signal to achieve the necessary discretional resolution.
The work referred to earlier by McComb and colleagues (2000) (see
Chapter 5) described how a single elephant, often the matriarch, can recognize
and identify 100 other elephants in the wider population. This ability calls upon
considerable neural capacity, including the ability to store and recall a large
amount of information.
Seyfarth and Cheney (2003b) as well as other psychologists maintain that
no animal other than humans calls to inform another animal or that the caller in-
tends to provide information to the receiver. Contrary to these assertions, current
work shows that contact calls between elephants allow the caller and receiver to
identify each other, resulting in an observed response such as each converging
on the other. Similarly, the “let’s go” rumble initiates a change in the behavior
of the receivers (i.e., an interpretation of the content of the call). Equally, accu-
mulating evidence in animal communication suggests that exchanges between
callers and receivers go well beyond the receiver acting as passive eavesdropper.
An example of a sequence of exchanges between a mother and her calf in the
Addo Elephant Park is described by Watson in his book Elephantoms (2003).
The mother and calf had been separated by a fence that was being repaired.
As related to Watson by a ranger who was at the scene, the calf responded to
inaudible instructions from the mother to move into cover (deep shade) about
20 yards on the other side of the fence. The calf did this and stood waiting mo-
tionless and almost invisible in the deep shade. The mother then rushed through
the gap in the fence and onto the road, confronting the fence repair crew in an
oncoming truck. In the turmoil of the charge and braking of the truck, both of
which generated clouds of dust, the calf slipped through the gap in the fence into
the dense bush of the park.
This event leaves us with at least two issues to ponder: Did the mother fig-
ure out the detailed strategy that led to the calf’s escape? Was the mother able
to communicate the essential details of this plan to her calf? The sequence of
events and deliberate actions and responses of each animal make chance a poor
candidate solution.
Masson and McCarthy ( When Elephants Weep, 1995, pp. 64–65) relate the
story of MaShwe, a work elephant, and her 3-month-old calf caught in the rising
flood waters of the Upper Taungdivin River in Burma (Myanmar). MaShwe was
able to keep her footing but the calf was in danger of being swept away. Despite
the mother’s efforts, the calf was swept downstream. MaShwe plunged after her
and managed to pin the calf against the bank with her head. She then lifted the
calf in her trunk, reared up on her hind legs, and placed her on a rocky ledge 5 ft
above the flood water. No sooner had MaShwe managed this when she fell back
and was swept away by the torrent.
76 Elephant Sense and Sensibility
Some half hour later as the elephant handlers, who had witnessed the entire
drama, were unsuccessfully trying to reach the calf, they heard a “defiant roar”
from MaShwe as she appeared on the opposite bank. As soon as she saw her calf
her calls turned to low rumbles, calming the calf.
She was not able, or was consciously unwilling, to cross the river until morn-
ing when the flood had begun to subside. MaShwe then crossed the river and
rescued the calf from the ledge.
Under these extreme conditions of stress, the mother was able to take coher-
ent action, respond to unexpected crises, return to the correct location, warn off
the humans who were getting involved, change her mode of communication to
calm the calf, keep the calf calm despite its precarious situation throughout the
whole night, and then under less dangerous conditions rescue the calf.
The entire sequence of events calls upon considerable rational behavior to-
part and an expected response by the calf. Describing the event in terms of
instinctual parental/offspring response would seem to be a poor alternative
explanation.
We, as humans, are justifiably impressed by the incredible complexity of our
own language and our ability to use it. Somewhat blinded by this fact, we have
failed to recognize the ability and complexity of language and communication
used by other species. Not knowing what is being said and how others are using
their language severely limits our ability to assess the capability of others and in
so doing we run the risk of ignoring them.
It is informative to reflect on what is known about the evolution of language
in humans. Most linguists anchor the origin of language to the emergence of
symbolic thought as it appeared for the first time in engravings and artifacts
crafted by early hominids. Geometric patterns engraved on an artifact found in
Blombos Cave, in the Cape Province of South Africa, is the earliest example
of such symbolic art. The Blombos artifact dates to only 70,000–80,000 years
ago (YA) (Figure 10.1) (McCarthy and Rubidge, 2005). At about 100,000 YA,
distinctive evolution in the vocal tract of early hominids took place. The FOX
P2 gene in the Hadzabe people of northern Tanzania appeared and aided in
speech formation and motor coordination. A “click” language may have been
developed and remains in evidence in the speech of the Hadzabe and San (South
Africa, Botswana, and Namibia) people. The click language is based on five
different clicks and consists of only about 140 distinct sounds. Most words are
only a single syllable and meaning is derived from the order of clicks and not
inflection. The characteristics of this early language in hominids are not that far
removed from the current status of language in elephants.
There are numerous reports of elephants drawing and painting, including
those in Thailand where the work is sold to tourists. Despite the possibility that
the Thai elephants may have received some form of instruction and certainly
receive rewards for their paintings, the paintings are nevertheless remarkable.
Mark Freeman’s demonstration seen in January 2014 of elephants painting in
Language Chapter | 10 77
FIGURE 10.1 An engraving in red ochre on a clay tablet from Blombos cave on the southern Cape coast. The specimen is 77,000 years old and provides the earliest eviden
ce of cognitive abilities central to modern human behavior. (Top) Clay tablet as it appears to the camera; (bottom) red ochre scratches. From McCarthy and Rubidge, 2005. Permission courtesy of Professor Christopher Henshilwood, University of Bergen, Norway.
Thailand must give the neuropsychologists considerable pause for thought. The
images painted are unquestionably those of elephants.
The execution in which successive brushstrokes made with apparent confi-
dence, combined to create a coherent composition, raises serious questions as to
the origin and conception of the ultimate product. Does the elephant have such a
completed image formulated in her mind? If so, such an achievement calls upon
some of the higher forms of neural processes. Observing the actual creation
of the image by the elephant strongly suggests that the equivalent of a mental
picture preexists. The initial stroke of the portrait recorded by Freeman was not
only elegant in composition and execution, but was clearly the foundation for
the ultimate creation.
Siri, a young Indian elephant, drew spontaneously in the dust of the floor of
her enclosure. She manipulated a pebble under her foot to scratch out designs
on the floor. She would then trace her design with the tip of her trunk, perhaps
detecting the smell of the exposed concrete or the residual flakes of the pebble.
Siri can draw with a pencil held in her trunk within the bounds of a 9 × 12 in
pad of paper held by her handler, Gucwa, on his lap. The fluid, flowing draw-
ings seen by artists not knowing who drew them were regarded as equal to and
78 Elephant Sense and Sensibility
often superior than those done by human artists. No rewards or prompting were
given to Siri either for her drawings on the concrete floor or on the pad of paper
(Masson and McCarthy, 1995, pp. 205–207).
As recent as the Blombos artifact is in evolutionary terms, it took nearly all
of the 80,000 years for humans to move from a spoken to a written language.
The Rosetta Stone dates from 196 b.c. with the possibility that Early Stone Age
art was in fact a form of written language. The pace of human development of
language since the appearance of the first written words is phenomenal. The
power now being unleashed even more rapidly by the expanding global capacity
to communicate can only be guessed at. This complexity of the human capac-
ity to communicate makes it even more difficult to recognize and interpret the