mise à jour du
23 novembre 2008
Teaching Dogs to Yawn, Sneeze, and Implications for Preparedness Theory and Observational Learning
Jacqueline R. Perkins
Veterinary Behaviour Consultant
Brisbane Queensland Australia
In: Kusonose, Ryo and Sato, Shusuke 39th Congress of the International Society for Applied Ethology, Kanagawa, Japan. 20-24 August, 2005


Tous les articles sur la contagion du bâillement
All articles about contagious yawning
Abstract In 1970 Dr Martin Seligman proposed psychological preparedness theory (On The Generality of the Laws of Learning: 1970 Psychological Review vol 77, pp. 406-418). Accordingly there are certain learning tasks which are contraprepared, or next to impossible. That a dog can not easily learn to yawn on cue was presented as such a contraprepared learning task. Juarbe-Diaz and Houpt (1996 p2) proposed that, because dogs are scent-dominated, they have a greater preparedness to learn via the stimulus of the citronella antibarking collar when compared with the static pulse electronic antibarking collar. This proposal is disputed. To date no-one has claimed to have taught a dog to yawn on cue easily, which has been taken by proponents of the theory as "proof". There is also the question of phylogeny: humans can yawn on cue, can dogs?
click to see the video
My group taught dogs to yawn on cue within a few trials. Another dog has been taught to sneeze on cue. The dogs are each part of a close multi-dog family household receiving a very large amount of human contact. Breed and age varies. Dogs 1,2 and 4 are a part of my own household, and represent every dog in that household. Dog 3 is a part of a co-workers household, which has three dogs in total. The time taken to2 train each dog varied from minutes to days. Each dog was easily trained to yawn on command using a combination of traditional gentle training methods. The primary method was yawning at the dog to solicit yawning which was then rewarded and put on cue. A hand signal was composed (a modification of the "speak" signal with the hand open but thumb and forefinger touching) and the yawn prompt rapidly faded. It is important to reward only full yawns from the outset. Dogs 2 and 4 learned by observing another dog yawn on cue for reward: dog 2 learned from dog 1 and dog 4 from dog 2. Furthermore, dogs 2 and especially 4 appeared motivated to compete using yawning to secure resources ahead of another dog.
It is proposed to limit Preparedness Theory to physiological preparedness, and that psychological preparedness either be revised (to exclude dogs yawning on cue) or abandoned: which depends on the theory of mind adopted and that is a subject for further research. Motivation and rapport with the trainer are more likely obstacles to teaching dogs social learning tasks and a lack of apparent learning does not imply those tasks are contraprepared. In a profoundly social animal such as the dog, many if not most behaviours are under social influence. Dogs may serve as a model for other species less able to accept humans into their social structure. Dogs may have greater cognitive abilities than previously believed. A close rapport with humans may be required for humans to measure their true cognitive abilities. Naïve dogs confined to a laboratory environment may be a confounding variable because it alters motivational status. The role of yawning within canine culture is elusive, but it may be used by pack leaders to synchronise group sleep. See accompanying video footage supporting my conclusions.
Yawning is a little studied puzzling behaviour performed by most vertebrates. It is a Modal Action Pattern and socially facilitated releasing stimulus (Provine 1996). A yawn follows an almost identical sequence in any individual, takes around 6 seconds and is easily recognised (Provine 1996). It includes slowly hyperextending the temporo-mandibular joint with a brief pause at hyper-extension, sometimes screwing up the eyes or closing them (sometimes causing the eyes to water), sometimes curling the tongue (seen in many dogs), sometimes includes a short rising vocalisation, is always accompanied by a short rapid 3 inhalation, then quickly returning the temporo-mandibular joint to a more physiological position and exhaling.
Just how voluntary is it across species, given that humans can yawn voluntarily? What purpose does it serve? To date studies indicate yawning may be positively correlated with ability to empathise (Platek, Critton, Myers, Gallup 2003), maleness in sexually dimorphic species (Provine 1996; Deputte 1994), comfort movement in carnivores and rodents (Deputte 1994), certain hormones (testosterone, oxytocin, ACTH) (Provine 1996), certain drugs (apomorphine, pilocarpine) (Provine 1996), maturity, dominance, boredom (Provine 1996), within an hour of sleeping or waking (Provine 1996), and as a transitional behaviour following social interactions (Deputte 1994). Interestingly, it has been found to not be primarily influenced by blood oxygen or carbon dioxide levels (Provine 1996). Yawning is associated with cholinergic and peptidergic excitation, and dopaminergic inhibition (Provine 1996).
Yawning has been claimed to serve various social functions including displacement behaviour (Provine 1996), and appeasement behaviour (Rugaas). There are social sanctions against public yawning in human society (Provine 1996 p181).
The hypothesis tested was that dogs can not be easily trained to yawn on command whereas, presumably, humans can yawn on cue easily. This hypothesis was presented by Martin Seligman in 1970 as a prime example of a psychological canine learning limitation. Learning limitations were hypothesised to occur because the dog's brain was said to be contra-prepared to easily learn certain tasks. The theory of psychological preparedness formed to explain this "phenomena". This theory has been accepted for many years and appears in most major texts. Juarbe-Diaz and Houpt use the theory to explain why they found citronella antibarking collars to be more effective than static pulse electronic antibarking collars (1967 p2).
The notion of psychological preparedness presented by Seligman apparently conflates anatomical, physiological and psychological preparedness. "…although the organism may have the necessary receptor and effector apparatus to deal with events, there is much variation in its ability to learn about relations 4 between events (Seligman p 409)." Seligman is saying that despite all circuitry and parts being present, if a behaviour is psychologically contraprepared then it takes numerous trials to teach if it can be taught at all.
With an additional 35 years of research behind us, we are now in a better position to understand that if all parts and circuitry are present, and the behaviour is difficult or impossible to teach, then some part of the circuitry may be missing ie Seligman's requirement that necessary receptor and effector apparatus being present, has not been met eg Seligman's example of psychological contrapreparedness that rats fail to learn by footshock (we now understand that at these levels of electrical shock, pain pathways dominate and learning pathways are suppressed, footshock alters the circuitry). This is an example of physiological contrapreparedness, not psychological contrapreparedness. All preparedness may be anatomical or physiological dependent on theory of mind. The theory of mind presumed influences the distinction between physiological and psychological preparedness. The theory of mind presumed herein is an epiphenomenon theory of mind in which behaviour and accompanying emotive states are not metaphysically identical to the underlying anatomy and physiology, hence psychological preparedness is distinct from physiological. This discussion warrants further consideration not undertaken herein.
Methods and Results
Together with two co-workers, in March 2003 we set out in an attempt to train dogs to yawn on command to question the validity of psychological preparedness theory. In total, we trained four dogs to yawn on command. They are each part of a close multi-dog family household receiving a very large amount of human contact. Breed and age vary. Dogs 1,2 and 4 are a part of my own household, and represent every dog in my household. Dog 3 is a part of a co-workers household, which has three dogs in total. The time taken to train each dog varied from minutes to days. Each dog was easily trained to yawn on command using a combination of traditional gentle training methods. A brief outline of each dog, unusual behaviours, and the method used follows (in order of training):
(1) "Dog1" a desexed male red with black/graying mask Staffordshire Terrier named Sargie. Dog1 came to live with my family in mid 2002 as a mature entire male exact age unknown. Dog1 had at least two previous homes and had aggression problems. He had mildly subluxating patellae and walked with straight hind legs dragging his hind nails. Dog1 received Rimadyl 50 mgs once daily long term. Dog1 had apparently not received even a basic education, and very slowly learned the house rules. Dog1 took over 3 months to train to sit reliably. A co-worker (my husband) developed an especially close relationship with Dog1. On the CBTI (Canine Behaviour Type Index) he was a SAM (spontaneous dominant low energy) personality type.
In March 2003, a co-worker began teaching Dog1 to yawn on command by yawning at dog1 and making it a game for him to yawn back. During the very first session dog1 yawned back 3 times in response to 7 or 8 human yawns directed at dog1 from a semi-crouch about 3 feet away. Dog1 was rewarded with verbal praise and games. Each yawn by dog1was precisely marked with "yes" previously trained-in as a reward marker. A hand signal was composed and introduced. The hand signal is a modification of the "speak" signal (open hand palm facing with thumb held apart from four fingers together, then close hand), but with the thumb and forefinger touching. The yawn prompt was faded over a few sessions. Dog1 was given a yawn training session early each morning for only a few minutes while on his special "place" or "mat" in the lounge room. After 2 sessions dog1 would pre-empt the first yawn cue and commence yawning and looking for his reward before the first cue, typical of any newly trained behaviour. Each yawn was marked with the word "yes" then rewarded with food immediately (initially dates, a favourite treat). Within a few trials over a few days Dog1 was yawning on cue reliably. Dog1 died in July 2004. He was not given a postmortem. The presumed cause of death was a bleeding hemangiosarcoma.
(2) "Dog2" a four year old desexed female red and white Staffordshire Terrier named Bladey. Dog2 was purchased from a breeder at 8 weeks of age. Dog2 learned to yawn on cue by observing dog1 receive food rewards for the action, marking the behaviour with rewards, plus shaping and approximation. The shaping and approximation were deemed to be a mistake. Dog 2 took longer to train to yawn on cue than the other 3 dogs giving many more mouth opens instead of yawns, apparently because the mouth opens had previously 6 been rewarded as part of the largely unsuccessful shaping procedure. Dog 2 is a very submissive dog and participates in the canine sport of Flyball. On the CBTI scale she is a SGH (spontaneous submissive high energy) personality type.
(3) "Dog3" a four year old desexed male camp-dog/dingo named Habari. Dog3 took just a few trials over a few days to train to yawn on cue, and was trained independently of dogs 1, 2 and 4 by another co-worker in a separate household. Dog3 belongs to a dog trainer veterinary nurse and is obedience trained weekly. Dog3 is dog aggressive and was rescued as a puppy from an aboriginal reserve. Dog3 had large scald burns on his back, heavy scarring persists. Dog3 was cued to yawn by his human yawning at him, then rewarded with food and praise. His human repeats the word "yawning" while yawning at dog3. On the CBTI scale he is an OBM (organised defiant low energy) personality type.
(4) "Dog4" a 20 month old desexed male German Shepherd Dog named Czar. Dog4 came to my home in July 2004. His full history is known. He was s successful show dog and went to stud but developed severe testiculitis and infertility resulted. He had been successfully obedience trialed at novice level qualifying twice. Dog4 is now used for companionship, personal protection, and participates in the canine sport of Flyball. Dog4 is very easy to train. Dog4 learned to yawn on cue in one trial within minutes by observing dog2 yawn for food rewards. Dog4 was new to the household and appeared keen to learn the new house rules and was seen to be observing dog2 closely. Dog4 lined up beside dog 2 when yawning for food, orientated towards the handler and yawned. This is a case of adult canine observational learning. Dog4 is not highly motivated by food but yawns for attention and petting. Dog4 can yawn the most repeatedly and reliably without fatiguing and with almost no approximating. Dog 4 took 6 months to train to fetch a tennis ball. The reason for this is assumed to be that the size of the tennis ball is below his prey threshold because dog 4 enthusiastically fetched a larger oblong shaped toy bone on the first trial. On the CBTI scale he is an OBH (organised defiant high energy) personality type.
(5) "Dog5" a 6 year old desexed female blue Australian Cattle Dog named Chase. Dog5 learned to sneeze on cue by the owner marking and rewarding the naturally occurring behaviour. Dog5 was first studied by me after trained to sneeze on cue. Her human is thought to recount the training procedure reliably and notes that dog5 sneezes exceptionally well when motivated by cheese, her favourite treat. Dog5 learned within 1- 3 trials. Dog5 belongs to a dog trainer who acquired dog5 from an animal shelter at 1 year of age. Dog5 displays fear and dominance aggression and had been returned to the animal shelter twice previously. Dog5 cocks her leg to urinate on vertical objects, has received extensive quantities of obedience training including in the canine sport of flyball. On the CBTI scale she is an OAH (organised dominant high energy) personality type.
Yawning may synchronise the transition into group sleep in social animals which explains why it has been positively correlated by other researchers with empathy and the leadership qualities of maleness and dominance; it is the leader's role to cue the group for sleep. It may additionally be used to "change the subject" (as a transitional behaviour) appearing as a displacement or an appeasing gesture in an attempt to alleviate social discomfort. Yawning research has been done on a limited number of species including rodents, primates and now dogs. Findings may not generalise to other species.
The findings herein indicate that dogs have the ability to empathise with their cohort as demonstrated by the intra and interspecial socially facilitated nature of yawning, in contrast with Anderson, Masako and Tetsuro who claim that "contagious yawning has only been reported in humans" (1st page)... they go on to demonstrate it in mature chimpanzees. Contagious yawning also occurs between humans and dogs.
All four dogs use the yawn in spontaneous attempts to solicit resources including human food, prized positions, and attention. It has become a tool for communication as do other learned behaviours. Dogs 1-4 in rank order of best at displaying yawning on cue (based on low number of: trials to acquire the behaviour; approximations when cued to yawn; and high reliability) are dog 4, 1 & 3, 2. This is in part explained by the theory that the high ranking male dogs are likely to be less socially inhibited to yawn on cue, if initiating yawning is indeed the leader's role. Dog2 and 3 by seem to be socially inhibited to yawn on cue (eg turns the head away). Like any acquired behaviour, its ease of acquisition is positively correlated with youth. Dog4 was the youngest dog and the best yawner.
Factors found to negatively impact canine ability to yawn on cue included commanding more than 3 yawns per session, high ambient temperature, presence of non-pack members, and various environmental distractions. After performing repeated yawn cueing, I personally found yawning to be inherently aversive. The Premack Principle explains that to increase the incidence of a low-incidence behavior (eg yawning), a high incidence behaviour should be used to reward it (eg a game or treat), rather than another yawn (which in this context becomes a punisher). It is best to end a yawn training session with a food treat and a rewarding activity after just one good yawn. Low incidence behaviours such as yawning and sneezing also pose the practical difficulty of training them to cue simply because it is difficult to catch the naturally occurring behaviour to mark and reward.
When dogs are panting due to a high ambient temperature, they are understandably reluctant to reduce their panting in order to yawn. It may be inhumane to expect a panting dog to yawn on cue hence yawn training sessions were regularly scheduled for 4-6 am in the subtropics. Also, the dogs were wide awake and keen to follow instruction at this time.
Positive factors when yawn training include using good behaviour principles, a shared social system, active and food rewards, conducting sessions on the dogs secure "place", and multiple dogs to incite competitive operant learning. The reward marker must follow the conclusion of the yawn as closely as possible without interrupting it. Treat the yawn as an all-or-nothing behaviour and reward only full convincing yawns from the beginning. Adopting the Behaviourist principle "you get what you pay for" seemed the most effective way to unambiguously communicate to the dog that a yawn was required and not a mouth open. Konorski (1967) apparently rewarded yawn approximations, or mouth opens, and reported that his dogs were truncating the yawn. Such is, however, not evidence for dogs' lack of ability to yawn on cue; the best training method of yawn training only rewards full yawns from the outset.
Dogs 2 and 4 learned to yawn on cue primarily through observational learning. They each observed another dog in the household receive rewards for yawning then likewise yawned for a reward. Dog 2 learned from observing dog 1, and dog 4 learned from observing dog 2. That they each orientated towards the trainer using hand signal only, and not the yawning dog, indicated it was a learned behaviour for reward and not merely a response to the releasing stimulus of a yawning dog. Dog 4 learned in a single trial and dog 2 learned within a few trials. According to Lindsay observational learning in adult dogs "remains in doubt, at least until more conclusive research is available. No experiment to date (that I know of) demonstrates observational learning in adult dogs." (Lindsay 2000 vol 1 p 272) Observational learning in adult dogs (dogs 2 and 4) has been demonstrated during training dogs to yawn on cue. Furthermore they actively compete with one-an-other during yawn training sessions for rewards thus improving their performance and enhancing the training process.
Dog 4 has refused to yawn on cue when near other unfamiliar dogs but obeyed when taken to a quieter more private corner of the park. It is unclear if this is simply a proofing issue or if there is some socially inhibiting factor/s involved eg yawning front-on near unfamiliar dogs may be a social faux pas. The coworker and owner of dog3 noticed an improvement in inter-dog relationships within the household since yawn training further reinforcing the idea that yawning has social significance within canine culture.
Dogs have complex motives and it would be premature to suggest any behaviour to be contraprepared without thorough research. Juarbe-Diaz and Houpt may be incorrect in suggesting that dogs are contraprepared to learn via the stimuli produced by static pulse electronic collars. Another explanation for the disappointing performance of the static pulse collar used in their study is that the programming of that device was too forgiving in that it shut off the delivery of the aversive stimulus if the dog continued to bark (p3). Thus the wearing dog was negatively rewarded for barking and it may be predicted that barking would increase. Reportedly the level of barking overall did decrease though perhaps the pattern of barking altered to reflect negative rewards for bursts of constant barking. It is noteworthy that the citronella collar 10 used to compare performance with the static pulse collar was not limited by the same "decisive feature" (p1) and could re-trigger each and every time the wearing dog barked.
Conclusion A close rapport with humans may be required for humans to measure dogs' true cognitive abilities. Thus dogs are a good experimental model because they are capable of forming a close rapport with humans to the point of sharing a social system. Social behaviours such as yawning seem to have social significance within canine culture. This factor can not be ignored when attempting to teach such behaviours. Perhaps the higher ranking dogs are more confident to perform social behaviours out of context even committing a canine social faux pas. The role of yawning may be to cue a transition into synchronised group sleep. As such it is a leader's role.
Adult dogs do learn by observation as evidenced by dogs 2 and 4 learning to yawn for reward through observing another dog rewarded for the action. Dogs may have greater cognitive abilities than previously believed. They can yawn and sneeze on cue, learn by observation, and Rico, a Border Collie in Germany appears capable of fast-mapping, that is forming a rough hypothesis about the meaning of a new word based on a single trial. Rico has a vocabulary of 200 and his abilities are comparable with a 2 year old child (Kaminski, Call, Fisher 2004).
Yawning on cue is not a point of difference between humans and dogs; mental time travel is a superior candidate (Suddendorf and Busby 2003). That dogs can yawn on cue suggests that other mammals, particularly carnivores, may also have the ability. Difficulties in demonstrating this may be due to barriers of a practical nature rather than in principle and dogs make a good model for overcoming some of these barriers due to their apparent ability to include humans in their social structure.
Preparedness theory needs to either be limited to anatomical and physiological preparedness, or if psychological preparedness is deemed to exist separately (depending on the theory of mind presumed) then the learning task of dogs yawning on cue needs to be added to the list of prepared behaviours for 11 physiological and/or psychological preparedness. Dogs' state of preparedness to learn via the stimuli produced by a static pulse electronic collar remains largely anecdotal with my own observations indicating they are prepared.
Anderson, Masako, Tetsuro 2004 Contagious Yawning in Chimpanzees Proceedings Royal Society London Anderson, Meno 2003 Psychological Influences on Yawning in Children Current Psychology 11, vol. 2
Deputte, B.L. 1994 Ethological study of yawning in primates.1. Quantitative analysis and study of causation in two species of old world monkeys Ethology 98, 221-245
Juarbe-Diaz, V. and Houpt, K.A. 1996 Journal of the American Animal Hospital Association
Kaminski, Call, Fischer 2004 Word Learning in a Domestic Dog: Evidence for "Fast Mapping" Science vol 304 pp1682-1683 12
Konorski, Robert 1996 Contagious Yawning and Laughter: Significance for Sensory Feature Detection, Motor Pattern Generation, Imitation, and the Evolution of Social Behaviour Social Learning in Animals: The Roots of Culture University of Maryland
Konorski 1967 Integrative Activity of the Brain Chicago, University of Chicago Press Lindsay, Steven R 2000 Handbook of Applied Dog Behaviour and Training vol 1 Iowa State Press Blackwell Publishing
Moyaho, Eguibar, Diaz 1995 Induced Grooming Transitions and open field behaviour differ in high-and low-yawning sublines of Sprague-Dawley rats Animal Behaviour, 50, 61-72
Perkins, Jacqueline & Dagley, Kenneth 2005 The Canine Behaviour Type Index 5th International Veterinary Behaviour Conference Proceedings American Veterinary Society of Animal Behaviour Purdue University Press
Platek, Critton, Myers, Gallup 2003 Contagious Yawning: the role of self-awareness and mental state attribution Cognitive Brain Research 17 pp223-227
Provine, Robert 1985 Yawning as a Stereotyped Action Pattern and Releasing Stimulus Ethology 72, pp109-122
Rugaas, Turid 2005 Calming Signals: The Art of Survival
Schiller, Francis Yawning? 2002 Journal of the History of the Neurosciences vol11 pp 392-401 13
Seligman, Martin 1970 On The Generality of the Laws of Learning Psychological Review vol 77, pp. 406- 418
Suddendorf, Busby 2003 Mental time travel in animals? Trends in cognitive science vol. 7 pp 391-396
Animal Minds 2004 New Scientist June pp 41-53

Le bâillement chez le chien - Yawning in dogs. Nathalie Tomczyk 2009
Teaching Dogs to Yawn, Sneeze, and Implications for Preparedness Theory and Observational Learning
JR. Perkins
Coping with fear and stress: licking and yawning
SR. Lindsay
Handbook of applied dog behavior and training 2000
Behavioural, saliva cortisol and heart rate responses to different types of stimuli in dogs
Bonne Beerda, Matthijs B. H. Schilder, Jan A. R. A. M. van Hooff, Hans W. de Vries and Jan A. Mol
Applied Animal Behaviour Science 1998;58(3-4):365-381.
Puppy behaviours when left home alone: a pilot study
Diane Frank, Michela Minero, Simona Cannas, Clara Palestrini 
Applied Animal Behaviour Science 2007;104(1-2):61-70.
Evaluation of the effects of a socialization program in a prison on behavior and pituitary-adrenal hormone levels of shelter dogs
Michael B. Hennessy, Angela Morris and Fran Linden
Applied Animal Behaviour Science 2006;99(1-2):157-17
Dogs catch human yawns
Joly-Mascheroni RM, Senju A, Sheperd AJ
Biology letters Animal Behaviour 2008;4(5):446-448
Adapting to the human world: Dogs' responsiveness to our social cues
Reid PJ.
Behavioural Processes 2008;80(3)325-333.
Comprehension of human pointing gestures in young human-reared wolves (Canis lupus) and dogs (Canis familiaris)
Virányi Z, Gácsi M, Kubinyi E, et al.
Anim Cogn 2008;11(3):373-387
Human-like social skills in dogs?
Hare B, Tomasello M.
Trends Cogn Sci 2005;9(9):439-444
The domestication of social cognition in dogs
Hare B, Brown M, Williamson C, et al
Science 2002;298(5598):1634-1636
Mouths wide open: yawning as a communicative behavior in dogs. Hoff AE 2001
Le bâillement chez le chien. Tomczyk N
Assessment of Stress in Laboratory Beagle Dogs Constrained by a Pavlov Sling
Stracke J, Bert B, Fink H, Böhner J.
Altex. 2011;28(4):317-325