Animal Cognition

, Volume 12, Issue 6, pp 833–837

Do dogs (Canis familiaris) show contagious yawning?


  • Aimee L. Harr
    • Department of PsychologyHiram College
  • Valerie R. Gilbert
    • Department of PsychologyHiram College
    • Department of PsychologyHiram College
Short Communication

DOI: 10.1007/s10071-009-0233-0

Cite this article as:
Harr, A.L., Gilbert, V.R. & Phillips, K.A. Anim Cogn (2009) 12: 833. doi:10.1007/s10071-009-0233-0


We report an experimental investigation into whether domesticated dogs display contagious yawning. Fifteen dogs were shown video clips of (1) humans and (2) dogs displaying yawns and open-mouth expressions (not yawns) to investigate whether dogs showed contagious yawning to either of these social stimuli. Only one dog performed significantly more yawns during or shortly after viewing yawning videos than to the open-mouth videos, and most of these yawns occurred to the human videos. No dogs showed significantly more yawning to the open-mouth videos (human or dog). The percentage of dogs showing contagious yawning was less than chimpanzees and humans showing this behavior, and considerably less than a recently published report investigating this behavior in dogs (Joly-Mascheroni et al. in Biol Lett 4:446–448, 2008).


DogsContagious yawningSocial contagionEmpathy


Viewing another individual yawn, reading about yawning, or even thinking about yawning can trigger contagious yawning in 40–60% of adult humans (Platek et al. 2003; Provine 1986, 1989). Investigations have shown that contagious yawning is not a phenomenon unique to humans, as chimpanzees (Anderson et al. 2004) and possibly stumptail macaques (Paukner and Anderson 2006) demonstrate contagious yawning effects, albeit to a lesser degree.

Current hypotheses attempting to explain contagious yawning have focused on its potential role in communication, social interaction, and empathy (Deputte 1994; Daquin et al. 2001; Gallese et al. 2004; Platek et al. 2005). While some consider contagious yawning to be an innate releasing mechanism (Provine 1986), others suggest that it has an empathetic basis and may be related to self-awareness (e.g., Platek et al. 2003). Support for this latter position comes from studies in humans correlating contagious yawning and scores on a test of empathy and self-recognition (Platek et al. 2003). Additionally, children with autism spectrum disorder show an impairment in contagious yawning (Senju et al. 2007). However, recent neuroimaging studies of contagious yawning have shown inconsistent results. Platek et al. (2005) noted activation in the posterior cingulate and precuneus, neural regions associated with self-processing (theory of mind and empathy). In contrast, Schürmann et al. (2005) reported activation in the superior temporal sulcus and not in the core regions of the human mirror-neuron system. The mirror-neuron system is believed to be an integral component of the neuronal framework for insight into other minds and empathy (Gallese et al. 2004).

Dogs are excellent species to address questions relating to social cognition. During domestication they were selected for social intelligence for their roles in human society (Cooper et al. 2003). Thus, dogs are predisposed to respond to the social cues of both humans and other dogs (Call et al. 2003; Hare and Tomasello 2005; Collier-Baker et al. 2004; Miklósi and Soproni 2006; Miklósi et al. 2003, 2007). Dogs might display a behavioral social contagion effect (James 1953; Scott and McRay 1967) in response to viewing either a dog or human yawn. A recent report investigating contagious yawning in dogs concluded that 72% of the dogs tested exhibited yawns in response to a human yawning (Joly-Mascheroni et al. 2008). Joly-Mascheroni et al. (2008) suggested that the heightened arousal that accompanies yawning (Daquin et al. 2001) may play a role in coordinating interactions and communication between dogs and humans. The present experiment further explored whether pet dogs would show contagious yawning. In particular, we tested whether dogs would show contagious yawning to both human and dog stimuli. Since dogs clearly respond to human and dog social signals, we hypothesized that dogs would display contagious yawning when presented with a stimulus of either an unfamiliar dog yawning or an unfamiliar human yawning.



Fifteen domestic dogs (Canis familiaris; six females and nine males) of various breeds and ages (1–11 years, M = 4.5, SD = 2.6; see Table 1) participated in this study. While the use of various breeds and ages was not optimal, such convenience sampling has been utilized in other published research on dog cognition (e.g., Adachi et al. 2007). Dogs were recruited for the study through flyer postings at local veterinary hospitals, post offices, and kennels. Informed consent was obtained from the dogs’ owners. To be eligible for participation, dogs had to be current on standard immunizations (e.g., rabies, distemper) and comfortable around strangers. All dogs had lived with the owner since they were puppies and had never participated in any behavioral experiments.
Table 1

Dogs’ breed and their age in this experiment


Number of subjects

Age(s) of dog in years

Bichon Frisé





2, 3, 3, 3, 4




Shepherd mix


6, 7, 11

Soft-Coated Wheaten Terrier



Standard poodle


2, 3, 4, 6

Experimental design

Clips of yawning and open-mouth facial expressions were recorded from three dogs (two females, one male) and three humans (two females, one male). The clips were then edited and sequenced to produce four videos: dog yawning, dog open-mouth, human yawning, and human open-mouth (see Fig. 1). Each video contained ten 3- to 5-s clips of the models yawning (or open-mouth) separated by 2 s of blank blue screen. Yawns recorded were naturally occurring and all four videos were matched for total length (1 min). In dogs, tension yawns can be distinguished from “sleepy” yawns as they are more intense and often associated with other signs of anxiety (e.g., whimpering, panting, or pacing). Any yawn classified as a tension yawn was noted and excluded from analysis. A 1-s orientation signal was used at the beginning of each video to orient the subject to the screen. With the exception of the orientation signal, the videos were shown silently using a HP Pavilion zd8000 laptop positioned directly in front of the subject.
Fig. 1

Frame sequences from the videos used in the experiment: a dog yawning, b dog open-mouth, c human yawning, d human open-mouth


Subjects were tested individually, and owners remained in a room separate from the testing site. Subjects were given a period of time to adapt to the new environment and experimenters before testing commenced. The subject then sat or lay down in front of the computer LCD monitor. One experimenter sat next to the subject, ensuring that the subject remained in front of the computer. The second experimenter sat behind the computer; testing began when the experimenter could observe the subject looking at the center of the screen. The four videos were shown with each video separated by a 3-min observation period. Order of yawn and open-mouth video presentation was counterbalanced across subjects. All yawns occurring during the videos and intervals between videos were recorded in real time by two experimenters who were present in the room. Interaction with subjects was kept to a minimum throughout the video screening and intervals. However, if necessary, one experimenter would encourage the subject to reorient to the video screen. When testing ended each dog was praised and given a food reward.


During the dog video stimuli, eight dogs yawned; during the human video stimuli, seven dogs yawned. All yawns recorded were judged to be “natural” yawns by both observers. Data were analyzed at both the group and individual level. Overall, yawning occurred at an average of 0.80 times (SD = 1.15) during and after the yawning videos (dog yawn video and human yawn video combined) and 0.73 (SD = 0.80) times during and after the open-mouth videos (dog open-mouth video and human open-mouth video combined). On average, dogs yawned 0.47 times (SD = 0.83) during and after the dog yawning video and 0.40 times (SD = 0.51) during and after the dog open-mouth video. Dogs yawned an equal amount of times after both human videos (M = 0.33, SD = 0.49). There were no significant effects at the group level, as determined by paired t tests.

One dog displayed a significantly higher frequency of yawns during yawning videos (total 5) than to the open-mouth videos (total 0) (binomial test, z = 2.27, N = 5, P = 0.011), and most of these yawns (4) occurred during the human video stimuli. This dog, a 2-year-old male retriever, was the sole dog in a household with no children. The owner indicated that the dog had a typical rearing environment and had received informal basic training. No dog displayed significantly more yawns to the open-mouth videos.


Higher frequencies of yawning in response to yawn stimuli (human or dog model) in comparison to open-mouth stimuli (human or dog model) were observed in one (6%) of the dogs tested. It should be acknowledged that this finding could be a chance occurrence. This percentage of dogs showing contagious yawning is quite less than the 33% of chimpanzees (Anderson et al. 2004) and 40–60% of humans (Provine 1986; Platek et al. 2003) that show a contagion effect of yawning, and is considerably less than the 72% of dogs reported to show a contagion effect in a recent study (Joly-Mascheroni et al. 2008). Interestingly, the subject who demonstrated the contagion effect in this study performed more yawns in response to viewing a human yawning than conspecifics’ yawns. This may reflect that dogs respond to the attentional state of humans (Call et al. 2003), as dogs use human signals such as gaze direction and distal pointing for informational purposes (Cooper et al. 2003). Dogs spend considerable time in human social experiences and are likely to be quite experienced with observing human facial gestures.

Although an experimental investigation into whether stumptail macaques display contagious yawning did yield positive results, the authors questioned whether the effect was due to tension or anxiety, as the yawns by the subjects were accompanied by self-directed scratching, a common indicator of tension in this species (Paukner and Anderson 2006). In the present study, we were careful not to record yawns displayed as a function of tension. We were attentive to whether a dog displayed any signs typically associated with anxiety; none of the subjects displayed such tension yawns.

The experiment reported here should be viewed as an initial investigation into the complex question of the degree to which pet dogs might show rudimentary elements of empathy. Our results indicate that in at least one dog, there was a clear increase in yawning after viewing an unfamiliar human or dog yawn. The percentage of dogs showing contagious yawning in this study is strikingly lower than the results reported by Joly-Mascheroni et al. (2008). Methodological differences likely explain the conflicting results between these two studies. While Joly-Mascheroni et al. (2008) used live human models, the present study used video clips of humans and dogs. Although the use of video stimuli could be considered advantageous as it standardized the procedure across all subjects and limited channels of communication transfer (Pongrácz et al. 2001), it is possible that subjects attended differently to the video rather than a live model. Differential reactivity of the human mirror-neuron system occurs when a subject observes a live or video motor act, with stronger reactivity seen when viewing a live motor act (Järveläinen et al. 2001). Research using video or LCD screens to present stimuli to dogs are limited, but successful studies have been conducted (Adachi et al. 2007; Pongrácz et al. 2003; Range et al. 2008; Rossi and Ades 2008). One study indicated that dogs successfully attended to human gestures on video (Pongrácz et al. 2003). To address this issue, experimental investigation testing the same subjects on both live and video models should be conducted. Unfortunately it was not possible to retest the current subjects with live human models. An additional reason that more dogs did not display a contagion effect may be that the models (both human and dog) used in this investigation were unfamiliar to the subjects. This may have decreased the likelihood of an effect, particularly in response to the stimuli of the human model. It is not unreasonable then to predict that the use of the dog’s owner or other familiar human in the video stimuli might have produced different results. There is a need for further experimentation on this issue, and unraveling whether such behavior in dogs reflects a rudimentary empathetic capacity will be an exciting area for future studies.


This research was supported by the Howard Hughes Medical Institute and Hiram College. We thank our human and dog models (especially Dr Chet Sherwood and Phoebe), and the dogs’ owners for their participation in the study. Experimental procedures were approved by Hiram College’s Institutional Animal Care and Use Committee and adhered to the Animal Behavior Society Guidelines for the Use of Animals in Research.

Copyright information

© Springer-Verlag 2009