Age, empathy, familiarity, domestication and call features enhance human perception of animal emotion expressions

Abstract

Vocalizations constitute an effective way to communicate both emotional arousal (bodily activation) and valence (negative/positive). There is strong evidence suggesting that the convergence of vocal expression of emotional arousal among animal species occurs, hence enabling cross-species perception of arousal, but it is not clear if the same is true for emotional valence. Here, we conducted a large online survey to test the ability of humans to perceive emotions in the contact calls of several wild and domestic ungulates produced in situations of known emotional arousal (previously validated using either heart rate or locomotion) and valence (validated based on the context of production and behavioural indicators of emotions). Participants (1024 respondents from 48 countries) were able to rate above chance levels the arousal level of vocalizations of three of the six ungulate species and the valence of four of them. Percentages of correct ratings did not differ a lot across species for arousal (49–59%), while they showed much more variation for valence (33–68%). Interestingly, several factors such as age, empathy, familiarity and specific features of the calls enhanced these scores. These findings suggest the existence of a shared emotional system across mammalian species, which is much more pronounced for arousal than valence.

Humans can identify reward-related call types of chickens

“…Whine calls were significantly less likely to be identified in a correct context-whines are wavering, high frequency tonal calls, which may not be perceived as easily as the harsh gakel calls. Greenall et al [26] tested the ability of humans to perceive emotions in the calls of ungulates produced in situations of known emotional arousal and valence and also identified a decrease in correct ratings with age over 20 years.…”

“…

Increases in arousal generally produce vocalizations that are harsher, louder, faster and longer, with a higher frequency ( f0) and a wider frequency range [15].Vocal correlates of valence are less easy to define, though positive contexts mainly elicit shorter call durations [16][17][18]. A growing body of studies has revealed that these acoustic properties appear to accurately predict human ratings of arousal and, to a lesser degree, valence in some animals, including dogs [19][20][21][22][23], pigs [24], silver foxes [25] and wild and domestic ungulates [26].In two studies, the success of participants in correctly classifying and describing emotionality in pig Sus scrofa and dog Canis familiaris calls, respectively, was attributed to use of Morton's motivationstructural rules [20,24]. These rules dictate that a call's frequency [27] is determined by the context it is produced in: high frequency calls are produced in fearful or appeasing contexts, whereas low frequency calls represent aggressive contexts.

…”

“…Vocal correlates of valence are less easy to define, though positive contexts mainly elicit shorter call durations [16][17][18]. A growing body of studies has revealed that these acoustic properties appear to accurately predict human ratings of arousal and, to a lesser degree, valence in some animals, including dogs [19][20][21][22][23], pigs [24], silver foxes [25] and wild and domestic ungulates [26].…”

A Relaxed Horse—A Relaxed Client? An Experimental Investigation of the Effects of Therapy Horses’ Stress on Clients’ Stress, Mood, and Anxiety

“…Two studies, focused on the interpretation of equine acoustic communicative signals (whinnies). Both studies showed that horse-experienced and horse-inexperienced participants were able to identify the valence and arousal of horses' vocalization above chance level [29,30]. Other studies focused on the human ability to identify equine affective states in horses' body language [31][32][33][34].…”

Pitch affects human (Homo sapiens) perception of emotional arousal from diverse animal calls.

“…In addition to vocalizer species, vocal F0 can be influenced by natural selection shaping signal structure, as well as vocalizer sex, body size, and age (Taylor & Reby, 2010), potentially reducing the accuracy of judgments made using the pitch rule. To speculate, this could contribute to the modest (though significantly above-chance) accuracy shown in several studies of heterospecific emotion perception, often 50%–60% on binary forced-choice tasks (Filippi et al, 2017; Greenall et al, 2022; Schwartz & Gouzoules, 2022). These considerations raise the question: how widely do humans apply the pitch rule, and to what extent does use of this heuristic depend on phylogeny and other factors?…”