160
D. Bo6et, J. Vauclair/Beha6ioural Brain Research 109 (2000) 143-165
vision, such as UV light [111], and offer false colour
representations for these subjects because the stimuli
are based on the trichromatic colour vision of humans
and not on the pentachromatic vision of birds (e.g.
[33]). Motion pictures are also matched for human
vision. In motion pictures, still images are shown in
succession to produce the impression of a continuous
moving image. But the frequency at which a flickering
stimulus starts to appear continuous is higher in birds
(notably in pigeons and in chicken) than in humans
[20,107]. Further, chickens are myopic (they can recog-
nise a conspecific only at a very short distance [18]) and
it is possible that they process far-distance views in a
manner distinct from that used for near-distance views.
The above considerations may thus explain some of the
problems encountered by birds in interpreting pictures
and why, for example, transfer is sometimes better for
black-and-white photographs than for colour photo-
graphs. Nevertheless, some of the experiments reported
earlier showed a good transfer from objects to pictures
or confusion between the slides or films and the animals
they represented.
Finally, a handful of studies have convincingly
demonstrated that responses to pictures are not limited
to birds and mammals; reptiles, fishes and even inverte-
brates reacted strongly to video images depicting bio-
logically significant stimuli (conspecifics, prey or
predators, for examples). It is, however, not surprising
that animals of different phyla respond to salient visual
cues in similar ways, i.e. as they would respond to real
objects; for many years, experimental ethology’s tech-
niques have used visual lures for identifying the stimu-
lus characteristics of social and aggressive behaviour in
animals (e.g. the pioneering and now classic studies
employing cardboard models of a Herring gull’s head
by Tinbergen and Perdeck, [95]).
6.2. Factors influencing picture recognition
The experiments reported in the section on acquired
responses to pictures fail to provide a clear and defini-
tive answer to the question of capacities of different
animal species for processing pictorial representation of
objects. Given that the most advanced and detailed
studies have been conducted with birds, it may be
useful to list some of the factors that authors have
highlighted as playing a determining role in the extent
and limits of picture recognition. Some of these factors
have been summarised in the discussion of d’Eath and
Dawkins’ [21] article reporting a failure of domestic
hens to discriminate between familiar and unfamiliar
conspecifics on videos and are thoroughly described in
the review by d’Eath [20]. A first and obvious factor has
to be mentioned: pictures, being still or in motion, are
abstractions from the reality they represent. Thus, even
if birds predominantly use visual cues to identify social
stimuli, auditory and olfactory information are also
present in encounters with real conspecifics but absent
in pictorial stimuli. Moreover, viewing a real conspe-
cific involves an interplay and exchange between the
pair of animals; this aspect of social interactions is also
lacking when the animals are presented with pictures.
Thus, because the picture cannot duplicate any interac-
tion between the viewer and the stimuli, a discrepancy
between the real object and the picture quickly becomes
apparent and renders the picture unrealistic and un-
challenging for the viewer.
Further, the experimental context of picture recogni-
tion presents subjects with stimuli which are reduced
along physical dimensions, such as the size (sometimes),
the colour and stereoscopic and motion parallax cues
used for perceiving depth. It should not be forgotten
that pictorial stimuli are abstract objects, which, be-
cause of their bi-dimensionality and because of the
other factors just mentioned, do not show all the visual
cues provided by real objects; such a reduction in the
informational content prevents the establishment of a
strict equivalence between real objects and their
representations.
A crucial factor in facilitating picture recognition is
related to the nature of the stimulus object; experiments
with animals have clearly demonstrated that categorisa-
tion of biologically relevant objects, such as food, is
readily generalised to photographs depicting the same
objects (see Ref. [105] for an example with pigeons and
[5] for an example with monkeys). Conspecifics and
predators also appear to be good stimuli for picture
recognition as shown by the evidence gathered with
birds [34], sheep [54] or macaques [14]. It is apparent
that interest for 2-D images, given their relative poverty
compared to real objects, can be reinforced if those
objects represent significant stimuli that belong to the
animal’s natural and/or social environment. To formu-
late the latter point in more general terms, it can be
stated that the value of the stimulus (e.g. familiarity
or/and experience with the to be recognised object) will
shape and facilitate discrimination and recognition
based on pictures (e.g. [48]).
Another factor, which seems obvious but is not al-
ways considered, is the complexity of the pictures pre-
sented, for example, very simple stimuli, like geometric
shapes or animals’ silhouettes (like those used by Cabe,
for example, [10]), might be more easily perceived than
complex photographs. This feature could explain why
the transfer observed in the preceding study was better
than that obtained with photographs of complex
scenes, such as those used by Dawkins et al. [19].
It is also important to note that the actual features of
the 2-D stimuli used by the subjects to solve a pictorial
task often remain unknown. In birds, such features can
be very different from those employed by humans. For
example, Troje et al. [99] tested pigeons in a categorisa-