haut de page























haut de page

mise à jour du 5 juin 2003
American Journal of Primatology
1988; 16; 165-169
Food-reinforced yawning in Macaca tonkeana
James R. Anderson, Daniel Wunderlich
Laboratoire de psychophysiologie, Université Louis Pasteur, Strashourg
Do chimpanzees yawn contagiously in response to 3d computer animations? Campbell MW et al 2008


Introduction : When adult male Old World monkeys yawn, lip retraction, the opening wide of the mouth, and the climactic tilting of the head upwards and backwards combine to expose the teeth clearly and to enhance the visual impact of the canines. Yawning is performed more frequently by adult males than by other age-sex classes, and its rate of occurrence is influenced by serum testosterone levels. Observations of increased rates of yawning in situations of tension or social conflict have given rise to the interpretation of yawning as a visual signal communicating an adult male's potential for aggressive intervention, with whether the yawn is directed toward an opponent being an important distinction

Given the apparent communicatory role of yawning in adult male Old World monkeys, the question arises as to whether yawns can be produced voluntarily. The widespread distinction between directed and nondirected yawns implies that the orientation of the act, at least, is controllable. Some data from observational studies may also be taken as indirect, albeit weak evidence for such control. For example, in olive baboons, the frequency of nondirected yawning in the proximity of other newly transferred males was positively correlated with the condition of the displaying male's canines. In Celebes macaques, rates of vawning by adult males varied with changes in their dominance status. There is,

however, little direct evidence. In the only previous experimental study on whether nonhuman primates can control their rate of yawning, Louboungou and Anderson obtained increased rates in two adult male pigtailed macaques through the use of food rewards. Extinction and recovery of yawning occurred when rewards were withdrawn and reinstated, respectively. These findings were striking and unexpected. In the present study we used adult males of another species of macaque, Macaca tonheana, to examine further the possibility of learned control over yawning, a behavior that Barbizet has described as "halfway between a reflex and an expressive movement."

Materials and methods :Subjects and Housing : The subjects were two group-reared, adult male tonkean macaques, Macaca tonkeana, recently removed from their group and living in individual indoor-outdoor cages. The subjects ("Nemo," 10.6 years old, and "Chou," 6.10 years old) were in visual contact with each other, but they interacted very little, either during experimental sessions or at other times. They were also in contact with their former group except during testing. Food (commercial monkey pellets) and water were available ad libitum.

Procedure : For all experimental sessions, each lasting 1 hour, the subjects were confined to the indoor part of their cages (1.9 X 1.4 x 2.0 m for Nemo, 2.9 x 1.4 x 2.0 m for Chou). The two experimenters each worked with one subject for all 22 sessions, which were conducted in ten phases over a 25-day period as follows: 1) Baseline (N = 2): spontaneous yawns were recorded; 2) Free rewards (N = 2): 20 small pieces of dried fig or fresh banana were given, distributed randomly throughout the session and independently of the subject's behavior; 3) Continuous reinforcement (CR) (N=4): All yawns produced by the subject were immediately reinforced by the experimenter, who gave the reward (a piece of dried fig or fresh banana) by hand; 4) Extinction EXT (N = 2): No rewards were given; 5) CR (N = 2); 6) EXT (N = 2); 7) CR (N = 2); 8) EXT (N =2); 9) Fixed ratio 3 (FR-3)(N = 2): a reward was given only after the subject produced three yawns; 10) EXT (N = 2). The subjects were always tested simultaneously.

Results : Figure 2 shows the results for each subject. To simplify presentation of the data, phases 1 and 2 have been combined to give BASE sessions, the CR phases 3, 5, and 7 have been combined, and EXT phases 4, 6, and 8 have been combined to give EXT 1. The data points in Figure 1 represent the average number of yawns produced in the four 15-minute blocks of each session.

During baseline sessions the subjects typically yawned between 2 and 4 times per 15-minute block, although Nemo yawned more frequently at the start of these sessions, apparently in response to being confined to the indoor cage section. Continuous reinforcement of yawning increased overall rates o the behavior in both subjects. This was especially evident in the first 15-minute block, with performance tapering off later in the session as the stibjects became satiated. However, in both subjects yawning rate in the last 15-minute block of CR sessions remained higher than in the corresponding block of baseline sessions.

The first 15-minute block of extinction sessions was marked by frequent yawning in both subjects. With the continued absence of reinforcement during extinction sessions there was an eventual return to baseline levels of yawning.

The most striking trends emerged during the FR-3 sessions. Chou averaged a peak of 16 yawns during the first 15-minute block in this condition. His yawning rate dropped to around 5 in the following two blocks (presumably because of satiation), then rose again during the final block. Nemo produced 15.5 and 17 yawns in the first two blocks respectively, 3.5 in the third block, and 6 in the final block.

In the final two extinction sessions, both subjects showed higher rates of yawning at the start of the session than at the end of FR-3 sessions. Both subjects then produced their lowest averages for any 15-minute block during these extinction sessions, although there was an unexplained increase (to 8.5 yawns) in the final block by Chou. There were no obvious behavioral signs of frustration (e.g., cage-shaking, scratching) during extinction.

Discussion : This study has shown that adult male Macaca tonkeana can control their yawning behavior to the extent that they will produce yawns to obtain food-rewards, and reduce their rate of yawning when food is no longer contingent upon the behavior. These results confirm the possibility of volitional control over yawning reported by Louboungou and Anderson for Macaca nemestrina However, whereas in that study the conditioned act in one subject was shaped through successive approximation, in the present study both subjects were rewarded from the start only for naturally occurring yawns. It is also noteworthy that the present positive results were obtained using

-1) prized tidbits as reinforcers, without food deprivation;
-2) relatively spacious caging conditions in which the subjects could move around and distance themselves from the experimenters. It seems likely that more traditional operant procedures including spatial restraint and food restriction would lead to even more clear-cut evidence for yawning as a conditionable response.

Steklis and Raleigh describe the important role of operant conditioning studies of communicative acts in the comparative approach to the neurobiology of communication in primates. Such research bas given convincing evidence for voluntary production of vocalizations in a range of primate species freviewed by Pierce, but there is no corresponding evidence regarding the production of facial expressions. Steklis and Raleigh point out that the low spontaneous rate of facial expressions presents a problem for operant paradigins, and Louboungou and Anderson felt that socially eliciting the to-be-reinforced facial expression created an emotional state in the subjects that interfered with learning. Satisfactory experimental paradigms for operant conditioning studies of primate facial expressions remain to be established, but in the meantime the positive results for yawning suggest that the attempt is worthwhile.

One aspect that was not covered in detail in the present study concerns the motoric components of the yawns produced. One subject (Chou) often seemed to struggle, mouth slightly open and rounded, to produce yawns, and sometimes abandoned the attempt. The other subject produced a mixture of "slow" and "fast" yawns, the latter resembling yawns produced when the subject was released into the outdoor cage section where he regained visual contact with his former group. Several authors have distinguished between yawns that occur in fatigue and those more readily identifiable as displays, positing differences in the extent of teeth-exposure, earflattening, and eyebrow-raising. A finer analysis of the components of yawns, possibly using videotape, could help clarify such differences, as well as the degree of conditionability of the two types. In any case, if some yawns funetion as a visual display, it would be useful for an individual to be able to control their production according to the situation. This ability has been demonstrated in the present study.


  1. Individually housed, adult male Macaca tonkeana learned to produce yawns for food rewards and eventually reduced their rates of yawning when rewards were no longer available (extinction).
  2. Adult male macaques therefore appear to be able to control yawning.
  3. It remains to be determined to what extent "true" and "display" yawns are distinguishable, and whether they are equally conditionable

« It is ironic that testosterone "the male sex hormone," is more closely associated with the yawning rate than with the mounting or intromitting rates » Charles Phoenix
Sexual steroids exert several effects on both central dopaminergic and oxytocinergic systems by acting either at the genomic or membrane level  
credit photo : "Asif A. Ghazanfar and Aristides Arrenberg"
Max Planck Institute for Biological Cybernetics
Tuebingen; Germany.