Anim Cogn (2008) 11:475–483
DOI 10.1007/s10071-008-0138-3
ORIGINAL PAPER
How do guide dogs of blind owners and pet dogs of sighted owners
(Canis familiaris) ask their owners for food?
Florence Gaunet
Received: 15 March 2007 / Revised: 3 January 2008 / Accepted: 17 January 2008 / Published online: 16 February 2008
© Springer-Verlag 2008
Abstract Although there are some indications that dogs
(Canis familiaris) use the eyes of humans as a cue during
human–dog interactions, the exact conditions under which
this holds true are unclear. Analysing whether the interactive modalities of guide dogs and pet dogs diVer when they
interact with their blind, and sighted owners, respectively,
is one way to tackle this problem; more speciWcally, it
allows examining the eVect of the visual status of the
owner. The interactive behaviours of dogs were recorded
when the dogs were prevented from accessing food that
they had previously learned to access. A novel audible
behaviour was observed: dogs licked their mouths sonorously. Data analyses showed that the guide dogs performed this behaviour longer and more frequently than the
pet dogs; seven of the nine guide dogs and two of the nine
pet dogs displayed this behaviour. However, gazing at the
container where the food was and gazing at the owner (with
or without sonorous mouth licking), gaze alternation
between the container and the owner, vocalisation and contact with the owner did not diVer between groups. Together,
the results suggest that there is no overall distinction
between guide and pet dogs in exploratory, learning and
motivational behaviours and in their understanding of their
owner’s attentional state, i.e. guide dogs do not understand
that their owner cannot see (them). However, results show
that guide dogs are subject to incidental learning and suggest that they supplemented their way to trigger their owners’ attention with a new distal cue.
F. Gaunet (&)
Laboratoire “Eco-Anthropologie et Ethnobiologie” UMR 5145,
Muséum National d’Histoire Naturelle,
CP 135, 57 rue Cuvier, 75231 Paris Cedex 05, France
e-mail: [email protected]
Keywords Human–dog interaction · Interactive
behaviours · Social cognition · Guide dogs · Socialisation
Introduction
Over the past few years, it has been shown that dogs can
read human social-communicative signals. For example,
dogs use the direction of attention of humans to obtain food
from the appropriate person and to perform forbidden
actions (e.g. Call et al. 2003; Virányi et al. 2004). They
understand human pointing, pointing and gazing, and gazing above chance level to locate hidden food or toys (e.g.
Hare et al. 1998; Miklósi et al. 1998; Agnetta et al. 2000;
Soproni et al. 2001, 2002). It has further been shown that
dogs can indicate to a human the location of a hidden target
(Miklósi et al. 2000). Together, these data conWrm folk psychology that there is an actual common space of communication between humans and dogs, based on perceiving the
direction of attention of the partner, triggering his/her attention and indicating the location of an object.
The present study focuses on dogs’ social-communicative signals for interacting with humans, and more speciWcally, with their owners. In an experiment by Miklósi et al.
(2000), in which the owner was naive about the location of
a hidden target (food or a toy located in one of three bowls
unreachable by the dogs), dogs displayed intentional communicative behaviours towards their owner. SpeciWcally,
when both the desired target and the owner were present,
the dogs tended to gaze at their owner and at the target, sniV
and lick their mouths, and alternately gaze at the target and
the owner. Vocalisations were also observed and were
always associated with gazing at the owner or at the location of the target (see also Hare et al. 1998, Study 3). These
signals were successfully used by the humans to locate the
123
476
hidden target, conWrming that they were attention-getting
and directional signals. Additionally, Miklósi et al. (2003)
found that the dogs looked longer at their caregiver than
wolves did when prevented from accessing food in a way
they had previously learned. The use of human gaze by
dogs as a social cue is thus a key to their interaction with
humans. These studies show that dogs have a propensity to
request help from humans using various types of signals
when they encounter an unsolvable problem (Cooper et al.
2003; Hare 2004), and that dogs use humans as tools (Hare
2004).
Convergent evolutionary and domestication processes
have been deemed to be involved in the expression of
human-like social skills in dogs (Hare and Tomasello 2005;
Hare et al. 2005; see also Miklósi and Soproni 2006 for a
discussion). The early, daily and long-standing closeness of
domesticated dogs to humans further provides dogs a special niche for the development of socio-communicative
skills (for a review, see Miklósi and Soproni 2006). Indeed,
because dogs learn to communicate and collaborate
throughout their lives with humans, they have ample opportunity to learn to use human behaviours, both intentional
and unintentional, to predict future events as a result of
simple learned contingencies (Cooper et al. 2003). Because
primary interactions between dogs and humans focus on the
management of a shared space (e.g. the human directs the
dog to come or to leave, the dog asks the human for food or
to go out, the dog or the human asks to play), communicative behaviours take place between the two species. Communicative behaviours are indeed viewed as a kind of
control system that allows group members to synchronize
activities and collaborate (Csányi 2000; see also Kerepesi
et al. 2005). Domestic dogs’ social-communicative skills
may thus be subject to learning, because they are exposed
to, and immersed in, the human world.
Importantly, dogs’ ability to use the most common
human deictic cues is not acquired (e.g. Hare et al. 1998;
Miklósi et al. 1998; Hare and Tomasello 1999; McKinley
and Sambrook 2000; Agnetta et al. 2000; Soproni et al.
2001, 2002). Even puppies with little prior human contact
use communicative signals such as gazing, and pointing
and gazing (Hare et al. 2002). However, Hare and Tomasello (1999) found that in a group of dogs ranging from
6 months to 4 years of age, only the youngest (a 6-monthold dog) was able to use a directional cue given by a conspeciWc, whereas the oldest dogs (two of 4-year-olds) used
only human cues. Miklósi et al. (1998) found that the dogs
learned to use bowing, nodding and head-turning, and one
dog learned to use directional glancing (see experiment 1);
the use of glancing improved signiWcantly in the group, but
the youngest dog never achieved the learning criteria (see
experiment 2). Soproni et al. (2001) also observed that the
ability to use eye direction as a cue developed across
123
Anim Cogn (2008) 11:475–483
sessions. In McKinley and Sambrook’s study (2000), two
dogs out of 11 succeeded in using eye direction; moreover,
trained gundogs were better at using arm pointing than
untrained gundogs. Consequently, inter-individual variability among dogs’ social-communicative and learning and
developmental processes may not be ruled out as factors in
the ability of dogs to use some human-given cues, especially
glancing.
Even if dogs seldom use human eye direction as a cue
for Wnding a toy or hidden food, a question that remains
unanswered concerns the exact conditions allowing the sensitivity of dogs to human eyes in referential communication, but also in perceiving the spatial locus of human
attention. For instance, when dogs are forced to choose
between two humans holding food, they prefer to beg from
a human whose head and eyes are visible and not covered
(Gácsi et al. 2004). Dogs are also more likely to avoid
approaching forbidden food when human eyes are open
than when they are closed (Call et al. 2003). Moreover,
dogs ignore a human’s gaze when the person stares into
space above the correct hiding location (Soproni et al.
2001) and they do not follow eye direction at all if there is
no attractive object (Agnetta et al. 2000). Finally, in the
presence of eVective and ineVective barriers, dogs make the
decision not to approach forbidden food when they cannot
see the human or when the human cannot see them (Bräuer
et al. 2004); dogs apparently project what the human can
see through a small window. Thus, although there are some
indications that human eyes, per se, are used by dogs as a
cue, the exact conditions under which this holds true
remain unknown. To address this issue, we chose a novel
approach (Cooper et al. 2003; Bering 2004; Tomasello and
Call 2004) to determine whether dogs use their owner’s
eyes as a cue for inferring the direction of attention of their
owner towards them. In the present study, guide dogs of
blind owners are compared with pet dogs of sighted owners. Guide dogs and pet dogs obtain diVerent lifetime experiences depending on the visual status of their owner. Our
goal was to determine whether dogs are sensitive to the
visual status of their owner, and whether, as a consequence,
dogs have somehow set up a way to trigger their owner’s
attention when the owner is blind.
The question of apprenticeship (Bering 2004) as the
result of enculturation or socialisation (Tomasello and Call
2004) among dogs is raised here. Socialisation is the sensitive period for the formation of primary social relationships
or attachments (Serpell 1995). On this basis, the two groups
of the present study were similarly socialised, as they were
reared in similar conditions at the earlier stage of their life
(i.e. in sighted families). However, they diVered in terms of
their adult living conditions. Guide dogs begin to get
“blind-socialised” later in life (after generally about 1.5–
2 years). Though some studies have shown the eVects of
Anim Cogn (2008) 11:475–483
early enculturation/socialisation on social-communicative
skills in apes (Bering 2004; Tomasello and Call 2004;
Miklósi and Soproni 2006), to date, no studies have compared groups of animals that were owned late in their life
by caregivers diVering in their perceptual skills, primarily
due to methodological issues (Bering 2004; Cooper et al.
2003). The present study is aimed to provide such data,
with a focus on the eVect of “blind-socialisation” among
dogs.
Guide dogs present a unique opportunity to study such
eVects on social-communicative behaviours in dogs: blind
owners and guide dogs have extensive interaction during
free time periods, and when synchronizing their actions
with humans during navigation, dogs have the ability to
interactively exchange roles with the human as the initiator
of actions (Naderi et al. 2001). Moreover sighted and blind
owners attend and respond to their dogs diVerently. As a
consequence, the daily interactions between guide dogs and
their blind owners could aVect the interspecies interactive
modalities of guide dogs, in a diVerent fashion from that of
pet dogs.
To test for any apprenticeship of social-communicative
cues, we adapted Miklósi et al.’s experiment (2003) and
measured exploratory and interactive behaviours during the
unsolvable test trial to get a broad overview of the dogs’
behavioural proWle. We expect that if a dog’s social cognition is attuned to the owner’s (non) visual status, then
sound and contact should replace to some extent the function of staring at the owner to trigger his/her attention. Consequently, less gazing at the owner and more sound
emissions and contacts in the guide dog group than in the
pet dog group should be observed. Further, guide dogs
should gaze less at the container where the unreachable
food is, and engage in less-sequential gazing between their
owners and the container, than pet dogs do.
Methods
Participants
One group of nine pet dogs of sighted owners and one
group of nine guide dogs of blind owners matched for
breed, age (t16 = 0.45, P = 0.66) and duration of the dyads’
common life (t16 = ¡0.25, P = 0.81) were involved in the
experiment (see Table 1 for the characteristics of the dogs).
Guide dogs were raised in a sighted family before receiving
their education between the ages of 6 and 18–24 months
and were transferred to the ownership of blind persons
when they were adults (1.83 years old on average). Five
guide dogs were from the same school and four other guide
dogs were from four other diVerent schools. On average,
pet dogs spent 2 h per day interacting with their owner
477
Table 1 Characteristics of the dogs (ages are in years)
Dog breed
Sex
Age
Number of years living
together (owner and dog)
Guide dog
Gold. £ Lab.
F
10
8
Golden
F
3
1
Labrador
F
3
1
Labrador
F
12
10
Labrador
M
5
3.5
Labrador
M
10
8.5
Labrador
F
5
3.5
Australian shepherd
M
7
5.5
Howavart
F
5.5
3
Gold. £ Lab.
F
7.5
7.1
Golden
M
7
2
Labrador
F
2.5
2.25
Labrador
M
8
7.75
Labrador
M
5.5
5.33
Labrador
F
7
5.5
Labrador
M
4
3.6
Pet dogs
Lab. £ shepherd
M
7
6.5
Golden
M
7
6.8
(e.g., play, educational games, obedience training, etc.), in
addition to morning, night and weekend walks and play.
Dog were not fed 4 h before the experiment. Owners
received a gratiWcation of 12 Euros for their participation.
Experimental settings
The dogs were tested in a large novel experimental room in
which the experimental area (3.88 m £ 2.70 m) was deWned (Fig. 1). For future analysis, the whole experiment was
videotaped by two cameras. Two identical 30-cm high plastic containers (30 cm in diameter) with two identical lids
covering the containers were used, as well as 17 pieces of
dry dog-food per dog. One of the two lids could be physically locked onto one of the containers by an invisible fastener, and a piece of food was hidden in this container just
before the training phase (cf. below).
Procedure
Before starting the experiment, the dogs were allowed to
explore the experimental room and area, including the two
lid-covered containers. The experimenter and owner discussed and agreed upon the verbal instruction that owners
would use to ask their dog to retrieve the food. The owners
were told to stand up during the experiment, and that during
123
478
Anim Cogn (2008) 11:475–483
(C1, C2, C3; C for “covered”). The experimenter then hid
behind the high fence (Fig. 1), asked the owner to remove the
piece of cloth from the dog’s eyes and to give the instruction. The dog was allowed to go and open the container on
its own.
80 cm high fence
Cont ainer
Experimental phase
Cont ainer
1.40 cm
h ig h w a l l
ner
Dog O(Dw
oorma t)
Ro
om
Ex
pe
r im
en
ta
la
re
a
80 cm
high fence
Experimenter
The diVerence between the last six trials and the Wnal one
(with the lid secured on the container, i.e. namely the
locked-container test trial) was that while the dog was
blindfolded, the experimenter exchanged the openable container with the locked one (hidden behind the 80 cm high
fence), and that the duration of the trial was 2 min and not
1.5 min.
The duration between the moment when the dog was
blindfolded and when it was unmasked was held consistent
throughout the 17 trials. At the end of each trial, the dog
returned to sit or stand close to its owner.
80 cm high fence 1.60 cm high fence
Fig. 1 Experimental arrangement
Data collection
Learning and motivation behaviours
(1) the Wrst ten identical trials, the experimenter would lift
the lid so that the dog could go and get the food, (2) the
next six trials lasting 1.5 min, the dog would have to lift the
lid by itself and (3) the last trial (2 min) involving the lid
secured to the container, the dog would try to remove the
lid by itself. The owners were asked not to move and interact with their dogs and to look ahead after giving the
instruction, especially during the last seven trials. They
were subjected to one or two training trials without their
dog to ensure they understood the task. The dogs were
unleashed.
Each of the 17 trials was started by asking the dog to sit
close to or to stand by its owner; the owner then covered its
eyes with a piece of cloth.
Training phase
For the Wrst ten trials, out of the dog’s view, the experimenter silently put one piece of food in the container, put
the lid on the top of the container and asked the owner to
remove the piece of cloth from the dog’s eyes. The owner
gave the instruction and the experimenter immediately
removed the lid after the dog had looked at her; the dog
was allowed to go and eat the food. Next, during the following six trials, the experimenter stopped demonstrating
the opening action: the food was hidden while the dog
was blindfolded; for the Wrst three trials the container was
half covered (HC1, HC2, HC3; HC for “half covered”)
and for the last three trials the container was fully covered
123
In order to assess any diVerence in the speed of learning
between the two groups of dogs during the block of six
learning trials (2 £ 3 trials), we measured:
• the latency to get the piece of food, which is the time
elapsed between the instruction given by the owner to the
dog and the time when the dog had the piece of food in
its mouth.
For the 2-min locked-container test trial, two variables were
measured to ensure that both groups were equally motivated to solve the task on their own, and two others ensured
equal food motivation:
• the latency to get to the container, which is the duration
between the owner’s instruction and the moment the dog
touched the edge of the lid with its nose to lift it.
• the number of attempts to lift the lid during the stays at
the container.
• the number of occurrences and duration of the dog staying at the container with its head bowed towards the container (a few cm) and sniVing it, or when it attempted to
lift the lid.
Behavioural observations during the 2-min locked-container test trial
The number and duration of occurrences of the following
behaviours were measured and submitted to inter-group
analyses.
Anim Cogn (2008) 11:475–483
• Exploratory behaviour: this behaviour refers to any
activity directed towards non-movable aspects of the
environment, including gazing at the ground with or
without sniVing, distal visual inspection (staring or scanning), and close visual or oral examination and looking
towards the non-visible experimenter.
The number of occurrences and duration of the following
interactive behaviours were measured:
• GazeOwner: the dog’s head/nose was oriented towards
the owner’s head.
• Vocalisations: the dog barked and/or whined.
• Contact: the dog pawed the owner, or touched him/her
with its nose or head or performed any form of bodily
contact.
• MouthLickingS: a noisy mouth-licking behaviour was
performed by the dog.
• MouthLickingNS: a silent mouth-licking behaviour was
displayed by the dog.
Combinations of some of the interactive behaviours were
also observed and recorded: GazeOwner + Vocalisation,
GazeOwner + MouthLickingS and GazeOwner + Contact.
Gazing at the container (GazeContainer) and gazing
at the container with sonorous mouth licking
(GazeContainer + MouthLickingS) were also observed and
recorded.
The sum of the durations of all 11 behaviours plus the
latency to get to the container and the duration of stay at
the container is 2 min for each dog.
Finally, the number of gazes at the owner followed
directly by a gaze at the container within 2 s, or vice versa
was recorded in both groups (GazeAlternation).
Two trained observers recorded the occurrences of
behaviours independently (one of them was naive with
respect to the aim of the experiment). In a few occurrences
(totalling 22 s out of the 120 s £ 18 dyads), the owner did
not follow the instruction to look ahead; the corresponding
dog’s behaviours were taken as exploratory behaviour.
Before data analyses, we assessed inter-observer agreements for the most frequent behaviours by means of parallel coding of 100% of the sample of durations. We
calculated Kappa coeYcients (Martin and Bateson 1986)
for durations and found relatively high values for Exploration: 0.81, GazeOwner: 0.83, Contact: 0.89, MouthLickingS: 0.85, GazeContainer: 1.00 and staying at the
container: 0.86.
We Wnally examined inter-group diVerences for the
latency for gazing at the owner and latency for the Wrst
interactive behaviour to be displayed (Wrst gaze at owner,
contact, vocalisation, sonorous mouth licking or combination of these behaviours).
479
Results
Learning and motivation behaviours
Non-parametric statistics were applied to the latencies to
get the piece of food (see Table 2 for medians). Mann–
Whitney tests did not reveal a group eVect for any of the trials (HC1: U = 27, P = 0.23; HC2: U = 38.5, P = 0.85; HC3:
U = 21, P = 0.08; C1: U = 36.5, P = 0.72; C2: U = 23.5,
P = 0.13; C3: U = 32.5, P = 0.47), with N guide dogs = N
pet dogs = 9. For each of the six trials, groups have thus
been pooled for the next analysis. Wilcoxon matched pairs
signed ranks tests revealed that the latencies for HC1&C1
pooled and HC2&C2 pooled did not diVer (n = 17,
T = 47.50, P = 0.16), nor did they diVer for HC2&C2
pooled and HC3&C3 pooled (n = 16, T = 49.0, P = 0.32);
however, the diVerence of latencies for HC1&C1 pooled
and HC3&C3 pooled was signiWcant, with HC1&C1 being
longer than HC3&C3 (n = 16, T = 30.50, P = 0.05) (see the
medians on the second part of Table 2).
Group comparisons for the four variables recorded during the 2-min locked-container test trial as an indication for
learning and motivation behaviours were carried out using
Mann–Whitney U tests, with N guide dogs = N pet
dogs = 9. The groups did not diVer for the latency to get to
the container (median = 1 for both groups; U = 40.50,
P = 0.99), for the number of attempts to lift the lid
(median = 3 for both groups; U = 34.00, P = 0.56), for the
number of occurrences of the dog staying at the container
(median = 2 for the guide dogs; median = 3 for the pet
dogs; U = 32.00, P = 0.45) and for the duration of the dog
staying at the container (median = 16 for the guide dogs;
median = 24 for the pet dogs; U = 20.50, P = 0.08).
These results show that the learning process does not
diVer between the groups of dogs and that the dogs were
equally motivated, both for attempting to solve the task on
their own, and for food in general.
Table 2 Medians of the latencies for getting the food according to the
dogs and trials
First set of trials
Second set of trials
HC1
HC2
HC3
C1
C2
C3
Guide dogs
3
3
2
5
5
4
Pet dogs
2
3
2
5
3
5
Two groups
pooled
3
3
2
5
4.5
4.5
Two groups pooled
HC1&C1
HC2&C2
HC3&C3
10
8
7.5
123
480
Anim Cogn (2008) 11:475–483
Behavioural observations during
the 2-min locked-container test trial
The guide dogs and the pet dogs explored the experimental
area for 61.29 and 61.57% of the time, respectively. The
number and duration of exploratory behaviours did not
diVer between the guide dogs (median = 20 for number;
median = 79 for duration) and the pet dogs (median = 19
for number; median = 72 for duration); see Table 3 for the
results of the Mann–Whitney analyses.
This test revealed that the guide dogs exhibited more and
longer instances of MouthLickingS (median = 3 for both
variables) than the pet dogs (median = 0 for both variables).
The number and duration of GazeOwner, Vocalisation,
Contact, MouthLickingNS, GazeOwner + Vocalisation,
GazeOwner + MouthLickingS and GazeOwner + Contact
did not diVer between the groups (Table 3; Fig. 2).
Table 3 also shows that the groups did not diVer in terms
of the number and duration of GazeContainer and
GazeContainer + MouthLickingS: the median of GazeContainer equals one for the guide dogs and the three other
medians equal two; for GazeContainer + MouthLickingS,
the median equals zero for both variables of both groups.
Further, the number of GazeAlternation did not diVer
between the guide and pet dogs (median = 0 for both variables).
Finally, it was found that the latency to gaze at the
owner did not signiWcantly diVer between the guide dogs
(median = 15) and the pet dogs (median = 29) (U = 15,
P = 0.13); two guide dogs and one pet dog did not look at
their owner at all. The latency for the Wrst interactive
behaviour (gaze at the owner, contact, vocalisation, sonorous mouth licking or combination of these behaviours) did
not diVer between the guide dogs (median = 15) and the pet
Table 3 Mann–Whitney tests
results for between-group analyses, with N guide dogs = N pet
dogs = 9
123
dogs (median = 27.5) (U = 25, P = 0.28). Among the guide
dogs, the Wrst interactive behaviour was to look at the
owner (for seven dogs) and to touch the owner (for two
dogs); among the pet dogs, six pet dogs looked at their
owner Wrst, one touched its owner Wrst, one vocalized Wrst,
and one did not gaze, touch, vocalize, mouth lick with
sound, or combine the two modalities at all.
Discussion
The aim of this study was to determine whether the behaviour of dogs, and more speciWcally the modalities of interaction of dogs, is attuned to the visual status of their
owners. We addressed this issue by comparing guide dogs
and pet dogs prevented from accessing food that they had
previously learned to access. Learning and motivational
behaviours did not diVer between the groups and both
groups explored the experimental area in a similar fashion.
These results conWrm that in a situation similar to a context
commonly experienced by dogs, working dogs such as
guide dogs display an exploratory repertoire with learning
and motivational behaviours that are similar to those of pet
dogs. Further, none of the main interactive behaviours (gazing at the owner, vocalisations, contact with the owner and
any combination of these behaviours), and neither the gaze
at the container nor alternate gaze, diVered between groups.
However, a new audible behaviour was observed, namely
sonorous mouth licks. These were signiWcantly greater in
number and duration in the guide dog group than in the pet
dog group. Cognitive innovations devised by animals in the
wild or in the human environment are indeed numerous
(Lestel 2004; Tomasello 2004; Bates and Byrne 2007). But
how may the behavioural pattern of the two groups be inter-
DiVerences between the guide and pet dogs
Duration
Number
U
P
U
P
Exploratory behaviour
37.00
0.75
37.50
0.79
GazeOwner
38.50
0.86
37.50
0.78
Vocalisation
36.00
0.32
36.00
0.32
Contact
24.00
0.08
24.00
0.08
MouthLickingS
13.00
0.009
14.00
0.012
MouthLickingNS
38.50
0.83
40.00
0.95
GazeOwner + Vocalisation
40.00
0.93
40.50
0.99
GazeOwner + MouthLickingS
31.50
0.14
31.50
0.14
GazeOwner + Contact
40.00
0.95
40.00
0.95
GazeContainer
39.00
0.89
40.00
0.96
GazeContainer + MouthLickingS
31.50
0.42
31.50
0.42
GazeAlternation
–
–
31.00
0.40
Anim Cogn (2008) 11:475–483
(a)
481
11
Guide dogs
Pet dogs
10
Number of behaviors
9
8
7
6
**
5
4
3
2
1
0
t
ner
tion ontac kingS ingNS ation kingS ntact
o
Ow alisa
c
k
lis
C
Lic
c
h
Lic Voca
hLi ner+C
t
o
u
V
uth
out
Mo
Mo wner+ erM+ zeOw
Ga
wn
zeO
Ga GazeO
ze
Ga
Duration of behaviors (sec.)
(b)
70
65
60
55
50
45
40
35
30
25
20
15
10
5
0
Guide dogs
Pet dogs
*
t
t
ner
tion ontac kingS ingNS sation kingS ontac
Ow alisa
li
c r+C
k
ic
C
i
e
a
c
L
i
L
z
c
c
h
L
o
ut
Ga
uth ne
Vo
uth
+V
Mo
Mo Ow
Mo wner
er+ Gaze
n
O
e
w
z
Ga
zeO
Ga
Fig. 2 The behavioural proWle of the dogs. a Number of behaviours
for the guide dogs and the pet dogs. b Duration of the interactive
behaviours for the guide dogs and the pet dogs. Non-parametric data
are represented as medians, and the box indicates the interquartile
range of 50% of the data. Whiskers extend to the smallest and largest
values and exclude outliers
preted, and what is the meaning of the sonorous mouth
licks?
At the outset, it is striking that gazing at the owner was
by far the most common interactive behaviour displayed
in both groups, and that this behaviour did not diVer
between groups. This conWrms that gazing at the owner is
a key factor in dogs’ social cognition when interacting
with humans (Miklósi et al. 2000, 2003), especially when
seeking help from the owner (Cooper et al. 2003; Hare
2004). This is further supported by the Wnding that the
latency to gaze at the owner (i.e. the Wrst gaze) did not diVer
between groups, and that the Wrst interactive modality was
gazing at the owner for the vast majority of the dogs in
both groups.
Second, the display of an increased number and duration
of sonorous mouth licks in the guide dog group may not be
as unexpected as it initially sounds, for several reasons.
First, the present task concerns food retrieval with the
owner present, which itself triggers mouth licking in dogs
(Miklósi et al. 2000). Second, touching the blind owner
would be too far from the referential object and the ears
replace the eyes in blind individuals for attending to space
and events (for a review of perceptual abilities of blind
individuals, see the introductory section of Gaunet and
BriVault 2005). Finally, guide dogs are taught neither to
beg and bark, nor to beg and whine.
The only directly available clue to tackle the meaning of
audible mouth licks and to interpret the diVerence between
the two groups is given by Miklósi et al. (2000). The
authors found that both mouth licking and sniVing behaviours signiWcantly increased when the dogs knew the location of a hidden and inaccessible piece of food and when
the owner—naive of that location—was present with the
dog. Since these authors did not observe these behaviours
when the dog was left alone with the hidden food, the emergence of these behaviours were attributed to the joint presence of the owner and food, and might indicate increased
levels of motivation because dogs associate the arrival of
the human with the retrieval of food (Miklósi et al. 2000).
There are indeed many opportunities for learning such signals as operant. For instance, noticing changes in the dog’s
behaviour, the owners may suppose that the dog might be
hungry, and “reward” this behaviour by providing food, or
they may respond to the directional gazing by dogs what
would cause dogs to conventionalize their gazing as a communicative exchange aimed at desired objects (Miklósi
et al. 2000). Both mouth licking—including a fortiori sonorous mouth licking—and staring at the owner may thus be
signals that draw the owner’s attention towards the dog
and/or a target (Fox 1971; Bradshaw and Nott 1995). These
signals may be a fortiori operant when both the owner and
the dog know that there is an inaccessible piece of food in
the room, though these signals do not provide a spatial reference to the target. These behaviours may, however, be
motivational signals per se, or signals aimed at triggering
the owner’s attention to a “desired” target.
Overall, because the groups did not diVer in terms of
gazing at the owner or gaze alternation, sonorous mouth
licking simply supplemented such behaviours rather than
replacing them, perhaps with the goal of capturing the
owner’s attention or to display motivation and/or interest.
Sonorous mouth licks would then be an addition, which
could be learned as a useful action to get help from the
owner. The fact that seven of nine guide dogs exhibited
sonorous mouth licks while only two pet dogs did, is never-
123
482
theless an argument that guide dogs may develop special
signals as a result of their owners’ diVerent attentional state.
Guide dogs may thus have supplemented a reduced reaction
of their owners to the visual signals they emit with a new
distal cue (or the greater number and the longer duration of
contacts with the owner—a proximal cue—in guide dogs
would have signiWcantly diVered from that of the pet dogs).
It might thus be that because blind owners cannot see the
dog trying to trigger his/her attention (e.g. by using the distal cue of staring) or see changes in the dog’s motivation,
dogs have incidentally developed an alternative tactic and
learned to use a new sound for triggering their owners’
attention from a distal location. Analysing whether this
behaviour would emerge when the target is inaccessible
and the owner is absent would tell whether this behaviour is
intentional.
Finally, because guide dogs were just as likely to gaze at
the owner, to gaze at the container and to exhibit gaze alternation as were the pet dogs, guide dogs were not actually
sensitive to the fact that their owners were not responding
to their gaze signals. The results thus suggest that guide
dogs did not understand their owner’s diVerent attentional
state (i.e. that their owners could not recognize the visual
signals dogs emit); if the dogs understood this, presumably
the guide dogs would have used MouthLickingS in place of
visual signals to gain their owner’s attention. Overall, and
in other words, the guide dogs did not detect that their owners could not see (them), but dogs have modiWed their interactive behaviours slightly, to adjust to their owners.
Further, the behavioural proWles of the dogs (Fig. 2) happened to superimpose fairly well in terms of the number
and duration of the interactive behaviours, for each group.
This shows that a constant duration of social-communicative behaviours is associated with the occurrence of each
behaviour, and further suggests that the behavioural modalities displayed when interacting with humans—once triggered—is regular in duration among the behaviours.
Taken together, the results suggest that the guide dogs’
behaviour is not attuned to the non-visual status of their
owner, but that sonorous mouth licking may supplement
staring at the owner to get the owner’s attention. The development of another attention-getting behaviour through
experience and the persistence of a generalized form of
gazing behaviour can be explained by a combination of reasons: Wrst, guide dogs grew up among sighted families and
still live surrounded by sighted people, and second, they
have an extensive experience of living with their blind
owners during their adulthood and are fed only by their
owner. The abilities of Philip, a 3-year-old male dog, who
was able to adjust his communicative behaviour to his disabled human partner to solve a cooperative problem-solving task (Topál et al. 2006), may also be the result of
Philip’s extensive experience with attending to his owner
123
Anim Cogn (2008) 11:475–483
and of the cognitive innovations that he has to perform regularly. Further, the development of another attention-getting behaviour through experience can also be explained by
the cooperative features of human–dog relationships and
the type of attachment of dogs to humans (Topál et al.
1998; Naderi et al. 2001, 2002; Kerepesi et al. 2005), and
by the sophisticated social-cognitive skills dogs posses (see
Hare and Tomasello 2005; Miklósi and Soproni 2006).
It must be emphasized that this is the Wrst study comparing social-communicative skills of groups of animals that
are owned late in their life by caregivers diVering in their
perceptual skills. This study is thus the Wrst evidence of
incidental learning in the adult dogs’ social-communicative
skills in an interspecies context, and the present evidence of
behavioural plasticity converges with Pongrácz-Rossi and
Ades’s (2008) study showing that an adult dog can learn to
use arbitrary signs to communicate requests to its owner.
To conclude, at this point, sonorous mouth licking may
be a motivational signal or a signal intended to trigger the
owner’s attention to a target (i.e. a communicative signal).
However, recent studies on vocalisations produced by dogs
and on the categorization by humans of these productions
have started to provide insights into the characteristics and
meanings of sound production in dogs. Yin and McCowan
(2004) analysed barking spectrograms in three diVerent contexts. A discriminant analysis revealed that dog barks are
divided into diVerent subtypes based on context. Moreover,
in another study, people were asked to describe the emotional content of several artiWcially assembled bark
sequences on the basis of Wve emotional states (aggressiveness, fear, despair, playfulness, happiness) (Pongrácz et al.
2007). The scoring of the emotional content of the bark
sequences was in accordance with the so-called Morton’s
structural–acoustic rules (e.g. low-pitched and atonal vocalisations signal aggressive intentions, in contrast to highpitched and tonal vocalisations signalling friendly/submissive intentions), suggesting that dog barks may present a
functional system for communication (see also Molnar et al.
2008). Studying the perception of sonorous mouth licking
by blind owners and the spontaneous reaction of blind owners to such sound may thus allow to clarify the role of this
behaviour in the context of human–dog interaction.
Acknowledgments This work was supported by Centre National de
la Recherche ScientiWque and conducted at the “Laboratoire Eco-Anthropologie et Ethnobiologie”, Muséum National d’Histoire Naturelle,
Paris, France. The experiments comply with the current laws in France
for animal and human research. Author thanks the guide and pet dogowner dyads for their interest and cooperation. Author is especially
grateful to B. Sauzeau and S. Reis for their contribution in the design
of the experiment and their help in determining behaviours to be collected, and to D. Sulinski for her contribution in the analysis of the videos. The author is also grateful to I. Guaitella for her comments on a
previous version of the manuscript and to J. Philbeck for his feed-backs
and corrections on the manuscript.
Anim Cogn (2008) 11:475–483
References
Agnetta B, Hare B, Tomasello M (2000) Cues to food locations that
domestic dogs (Canis familiaris) of diVerent ages do and do not
use. Anim Cogn 3:107–112
Bates LA, Byrne RW (2007) Creative or created: using anecdotes to
investigate animal cognition. Methods 42(1):12–21
Bering JM (2004) A critical review of the ‘enculturation hypothesis’:
the eVects of human rearing on great ape social cognition. Anim
Cogn 7:201–213
Bradshaw JWS, Nott HMR (1995) Social and communication behaviour of companion dogs. In: Serpell JA (ed) The domestic dog: the
biology of its behaviour. Cambridge University Press, Cambridge, pp 116–130
Bräuer J, Call J, Tomasello M (2004) Visual perspective-taking in dogs
(Canis familiaris) in the presence of barriers. Appl Anim Behav
Sci 88:299–317
Call J, Bräuer J, Kaminski J, Tomasello M (2003) Domestic dogs are
sensitive to the attentional state of humans. J Comp Psychol
117:257–263
Cooper JJ, Ashton C, Bishop S, West R, Mills DS, Young RJ (2003)
Clever hounds: social cognition in the domestic dog (Canis familiaris). Appl Anim Behav Sci 81:229–244
Csányi V (2000) The ‘human behaviour complex’ and the compulsion of
communication: key factors in human evolution. Semiot 128:45–60
Fox MW (1971) Behaviour of wolves, dogs and canids. Jonathan Cape,
London
Gácsi M, Miklósi Á, Varga O, Topál J, Csányi V (2004) Are readers of
our face readers of our minds? Dogs (Canis familiaris) show situation-dependent recognition of human’s attention. Anim Cogn
7:144–153
Gaunet F, BriVault X (2005) Exploring the functional speciWcations of
a localized wayWnding verbal aid for blind pedestrians: simple
and structured urban areas. Hum Comput Interact 20(3):267–314
Hare B (2004) Domestic dog use humans as tools. In: BekoV M (ed)
Encyclopedia of animal behavior, vol 1. Greenwood Press, Westport, pp 277–285
Hare B, Tomasello M (1999) Domestic dogs (Canis familiaris) use human and conspeciWc social cues to locate hidden food. J Comp
Psychol 113:173–177
Hare B, Tomasello M (2005) Human-like social skills in dogs? Trends
Cogn Sci 9:439–444
Hare B, Call J, Tomasello M (1998) Communication and food location
between human and dogs (Canis familiaris). Evol Commun
2:137–159
Hare B, Brown M, Williamson C, Tomasello M (2002) The domestication of social cognition in dogs. Science 298:1634–1636
Hare B, Plyusnina I, Ignacio N, Schepina O, Stepika A, Wrangham
R, Trut L (2005) Social cognitive evolution in captive foxed is
a correlated by product of experimental domestication. Curr
Biol 15:226–230
Kerepesi A, Jonsson GK, Miklósi Á, Topál J, Csányi V, Magnusson
MS (2005) Detection of long-term temporal patterns in dog–
human interaction. J Behav Proc 70:69–79
483
Lestel D (2004) L’animal singulier. Seuil, Paris
Martin P, Bateson P (1986) Measuring behaviour. Cambridge University Press, Cambridge
McKinley J, Sambrook TD (2000) Use of human given cues by domestic dogs (Canis familiaris) and horses (Equus caballus). Anim
Cogn 3:13–22
Miklósi Á, Soproni K (2006) A comparative analysis of animals’
understanding of the human pointing gesture. Anim Cogn
9:81–93
Miklósi Á, Polgárdi R, Topál J, Csányi V (1998) Use of experimenter
given cues in dogs. Anim Cogn 1:113–121
Miklósi Á, Polgárdi R, Topál J, Csányi V (2000) Intentional behaviour
in dog–human communication: an experimental analysis of showing behaviour in the dog. Anim Cogn 3:159–166
Miklósi Á, Kubinyi E, Topál J, Gasci M, Viranyi Z, Csányi V (2003)
A simple reason for a big diVerence: wolves do not look back at
humans but dogs do. Curr Biol 13:763–766
Molnar C, Kaplan F, Roy P, Pachet F, Pongrácz P, Dóka A, Miklósi Á
(2008) ClassiWcation of dog barks: a machine learning approach.
Anim Cogn (in press)
Naderi Sz, Miklósi Á, Dóka A, Csányi V (2001) Co-operative interactions between blind persons and their dogs. Appl Anim Behav Sci
74:59–80
Naderi Sz, Miklósi Á, Dóka A, Csányi V (2002) Does dog–human
attachment aVect their interspeciWc cooperation? Acta Biol Hung
53:537–550
Pongrácz-Rossi A, Ades C (2008) A dog at the keyboard: using arbitrary signs to communicate requests. Anim Cogn (in press)
Pongrácz P, Molnár Cs, Miklósi Á, Csányi V (2007) Acoustic parameters of dog barks carry emotional information for humans. Appl
Anim Behav Sci 100:228–240
Serpell JE (1995) The domestic dog: its evolution, behavior and interactions with people. Cambridge University Press, Cambridge
Soproni K, Miklósi Á, Topál J, Csányi V (2001) Comprehension of human communicative signs in pet dogs (Canis familiaris). J Comp
Psychol 115:122–126
Soproni K, Miklósi Á, Topál J, Csányi V (2002) Dogs’ responsiveness
to human pointing gestures. J Comp Psychol 116:27–34
Tomasello M (2004) Aux origines de la cognition humaine. Retz: Paris
(p10). Translated from Tomasello M (1999) The cultural origins
of human cognition. Harvard University Press, Cambridge
Tomasello M, Call J (2004) The role of humans in the cognitive development of apes revisited. Anim Cogn 7:213–215
Topál J, Miklósi Á, Dóka A, Csányi V (1998) Attachment behaviour
in the dogs: a new application of the Ainsworth’s strange situation
test. J Comp Psychol 112:219–229
Topál J, Erdõhegyi Á, Mányik R, Miklósi Á (2006) Mindreading in a
dog: an adaptation of a primate ‘mental attribution’ study. Int J
Psychol Psychol Therap 6(3):365–379
Virányi Z, Topál J, Gácsi M, Miklósi Á, Csányi V (2004) Dogs respond appropriately to cues of humans’ attentional focus. Behav
Process 66:161–172
Yin S, McCowan B (2004) Barking in domestic dogs: context speciWcity and individual identiWcation. Anim Behav 68:343–355
123