Scientists have finally learned to speak pig, thanks to a new algorithm that decodes the animals’ vocalizations to reveal their emotional state. Presenting their porcine paraphraser in the journal Science Advances, the study authors explain how they trained an artificial intelligence system using 7,414 recordings from 411 pigs in order to determine which grunts, squeals, and oinks convey happy feelings and which express negative emotions.
"With this study, we demonstrate that animal sounds provide great insight into their emotions,” said study author Associate Professor Elodie Briefer in a statement. “We also prove that an algorithm can be used to decode and understand the emotions of pigs, which is an important step towards improved animal welfare for livestock."
Like all mammals, pigs possess numerous limbic structures within their brains that are heavily associated with emotion. However, the extent to which animals are able to experience complex emotional states remains a hotly debated topic.
By seeking to answer this question, scientists hope to enhance farmers’ capacity to care for their livestock, specifically by keeping tabs on their emotional health. To that end, the study authors recorded the noises made by commercial pigs of different ages in a range of situations eliciting both positive and negative emotions.
For example, the sounds made by piglets when suckling or by pigs that were reunited with family members following a period of separation were recorded as examples of happy vocalizations. In contrast, emotionally negative situations included castration, fights, and slaughter.
Recordings were taken from 19 different contextual situations representing the full range of scenarios encountered by farm pigs throughout their life cycle. Some of these took place in experimental stables, where animals were presented with a series of novel experiences such as the opportunity to interact with unfamiliar objects.
Previous research has indicated that piglets tend to let out more high-pitched calls when they are distressed, and the results of this study appear to corroborate this aspect of pig speak. In general, higher-pitched squeals and screams were heard most often during negative situations while low-frequency grunts occurred during both negative and positive scenarios.
All calls, whether low or high, tended to be of a shorter duration when pigs were content, while negative emotions were typically associated with vocalizations that fluctuated in volume.
"There are clear differences in pig calls when we look at positive and negative situations,” said Briefer. “In the positive situations, the calls are far shorter, with minor fluctuations in amplitude. Grunts, more specifically, begin high and gradually go lower in frequency.”
“By training an algorithm to recognize these sounds, we can classify 92 percent of the calls to the correct emotion."
Having successfully decoded the emotional content of pigs’ grunts, Briefer says that her team’s algorithm could now be used to create a smartphone app to allow farmers to interpret the noises made by their animals.