Abstract
Past research has focused on the ability of domestic dogs to recognise human communicative cues such as human pointing, eye gaze, attachment behaviour, social learning, and responses to affective stimuli. However, there is still much to be learned about dogs’ ability to recognise human emotions. The main objective of this study was to investigate whether domestic dogs understood the emotions expressed by a human experimenter. I examined four emotions (anger, sadness, fear, happiness) and three types of expressions (dynamic, unimodal, bimodal) in four studies. The four studies covered the following: (1) dogs’ response to commands accompanied by dynamic expressions of human happiness and anger; (2) dogs’ social referencing of human expressions of fear and happiness, and one control “confused” expression; (3) dogs’ and human infants’ preferential looking to angry, happy and sad human faces when listening to a matching emotion voice. In addition, infants’ facial expressions were analysed to examine whether their facial expressions differed as a result of the affective displays they viewed, and one additional control experiment was conducted to examine dogs’ ability to perform gender-matching of human faces to voices, and (4) dogs’ and young adult humans’ physiological (cortisol levels) and behavioural response to human infant crying, human infant babbling, and white noise. In Study 1, I found that dogs have differential responses to the experimenter’s happy and angry expressions emanating from the body, face or voice cues during the emoting phase, but not when left alone with food. In Study 2, dogs responded differentially to happiness and fearful expressions, but had similar responses to both fearful and control conditions, suggesting that they might not have understood the fearful expression meaningfully. Results from Study 3 showed that both dogs and infants looked less at sad faces (irrespective of the matching voice), and had no preference for looking at either happy or angry faces. Also, infants displayed a sad expression when viewing a sad face. For the gender-matching task, dogs were able to match male faces to voices, but not for female stimuli. Finally in Study 4, I found that both dogs and humans had increased cortisol levels after listening to a human infant crying but not to babbling or white noise. Dogs also showed a combination of alert and submissive behaviour when listening to crying. In conclusion, the results provided some evidence that dogs tended to respond differently to human emotional expressions and, similar to humans, may have an aversion to human expressions of sadness, indicating the presence of emotional contagion, a form of rudimentary empathy.