Pig Grunts Can Reveal Sadness or Happiness

It’s easy to tell the difference between a dog’s excited, happy bark and an angry, warning growl. But not all animals are as easy to decode.

Researchers recently used thousands of recordings to decipher the different grunts that pigs make. They created an algorithm that helped them determine whether pigs were experiencing a positive emotion, a negative emotion, or something in between.

“We knew that pig calls can indicate their emotions, but most previous work had been carried out on a specific age class (e.g. piglets, or weaned pigs) and call type (e.g. only grunt, or only screams),” co-author Elodie Briefer, an associate professor in the University of Copenhagen’s Department of Biology, tells Treehugger.

“Here, many teams working on vocal expression of emotions in pigs gathered together to look, more widely, at the potential for building a tool that could recognize pig emotions from their vocalizations in a farm setting, meaning across age class, conditions, and call types.”

For their study, the team collected new and existing recordings from commercial facilities including research farms and slaughterhouses. They analyzed more than 7,400 calls from 411 pigs in 19 different contexts ranging from birth to death.

The results were published in the journal Scientific Reports.

Positive and Negative Emotions

The researchers used recordings from both positive and negative situations. The positive scenarios include when piglets nurse with their mothers or when separated family members are reunited. The negative situations include separation from family members, fights between pigs, and castration.

The researchers also created situations that were somewhere in the middle of those two extremes. They made an area with toys or food and a similar area without any of those positive factors. They also put unfamiliar objects into the area for the pigs.

They studied 7,414 recordings, analyzing them to look for a pattern between sounds and emotions and to see if they could tell the difference between the positive and negative situations based on the calls. As other studies suggested, these researchers collected more high-frequency calls such as squeals and screams in negative situations. Low-frequency calls such as grunts and barks were found in both positive and negative scenarios.

“Positive calls (whether they are low [e.g. grunt and barks] or high [screams and squeals] call types) are shorter in positive than negative situations and have less amplitude modulations,” says Briefer. “Other parameters were changing also but in a different direction depending on whether we look at low or high frequency calls.”

The researchers trained a machine learning algorithm to recognize whether a sound is associated with a positive or negative emotion. The algorithm was able to classify the sounds to the correct corresponding emotions with an accuracy rate of 92%.

Focus on Mental Health

Most animal lovers may not be surprised by these findings, and neither were the researchers. 

“We knew that pig vocalizations indicate their emotions, so that wasn’t surprising, but we were rather amazed at the high accuracy that we can reach with an automated algorithm,” Briefer says.

The researchers hope these findings can be used to gauge and improve the health of livestock. While farmers know that it’s important to maintain the physical health of their animals, this puts the focus on mental health, researchers say.

“It could be implemented in a tool that could record groups of pigs and warn the farmer when a certain threshold of negative calls has been reached for example, allowing the farmer to go check on the group of animals,” Briefer says. “It could also let the farmer know when the pigs seem 'happy,' so when there are many positive calls produced. Farmers could thus use this tool to know if changes they implemented in the barn, for example, results in good welfare for the animals or not.”